Depression prevalence using the HADS-D compared to SCID major depression classification: An individual participant data meta-analysis

https://doi.org/10.1016/j.jpsychores.2020.110256Get rights and content

Highlights

  • No studies have compared HADS-D results to prevalence from diagnostic interviews.

  • We evaluated data from 41 studies (6005 participants, 689 with major depression).

  • Prevalence based on HADS-D ≥ 8 (24.5%) was more than double the SCID (11.6%).

  • HADS-D ≥ 11 (10.7%) was similar but single-study differences highly heterogeneous.

  • The HADS-D should not be used as a substitute for a validated diagnostic interview.

Abstract

Objectives

Validated diagnostic interviews are required to classify depression status and estimate prevalence of disorder, but screening tools are often used instead. We used individual participant data meta-analysis to compare prevalence based on standard Hospital Anxiety and Depression Scale – depression subscale (HADS-D) cutoffs of ≥8 and ≥11 versus Structured Clinical Interview for DSM (SCID) major depression and determined if an alternative HADS-D cutoff could more accurately estimate prevalence.

Methods

We searched Medline, Medline In-Process & Other Non-Indexed Citations via Ovid, PsycINFO, and Web of Science (inception-July 11, 2016) for studies comparing HADS-D scores to SCID major depression status. Pooled prevalence and pooled differences in prevalence for HADS-D cutoffs versus SCID major depression were estimated.

Results

6005 participants (689 SCID major depression cases) from 41 primary studies were included. Pooled prevalence was 24.5% (95% Confidence Interval (CI): 20.5%, 29.0%) for HADS-D ≥8, 10.7% (95% CI: 8.3%, 13.8%) for HADS-D ≥11, and 11.6% (95% CI: 9.2%, 14.6%) for SCID major depression. HADS-D ≥11 was closest to SCID major depression prevalence, but the 95% prediction interval for the difference that could be expected for HADS-D ≥11 versus SCID in a new study was −21.1% to 19.5%.

Conclusions

HADS-D ≥8 substantially overestimates depression prevalence. Of all possible cutoff thresholds, HADS-D ≥11 was closest to the SCID, but there was substantial heterogeneity in the difference between HADS-D ≥11 and SCID-based estimates. HADS-D should not be used as a substitute for a validated diagnostic interview.

Introduction

Accurately measuring depression prevalence in different populations is important to understand disease burden, interpret research on etiology, and utilize healthcare resources as efficiently as possible [1]. In mental health research, diagnostic interviews are required for diagnosis of major depression [2,3]. These interviews, however, are costly to administer, especially in large groups, due to the time and trained personnel required to conduct them properly. Therefore, self-report screening questionnaires are sometimes used as an inexpensive alternative to evaluate depression prevalence, with the percentage of patients scoring above a cutoff threshold being described as the prevalence of depression [4,5]. Screening tool cutoffs, however, are typically set to cast a wide net and identify many more individuals for further assessment than will meet diagnostic criteria. Thus, commonly used screening tools tend to overestimate depression prevalence, sometimes substantially [5].

A previous study used an individual participant data meta-analysis (IPDMA) approach to compare prevalence based on a depression screening tool with prevalence based on a validated diagnostic interview. That meta-analysis examined prevalence based on the Patient Health Questionnaire-9 (PHQ-9) using the standard cutoff of ≥10 compared to prevalence based on the Structured Clinical Interview for the DSM (SCID) among 9242 participants from 44 primary studies [6]. Compared to the SCID, PHQ-9 ≥10 overestimated prevalence by 11.9%; across included studies, the mean and median ratio of PHQ-9 prevalence to SCID-based prevalence were 2.5 and 1.9. In that study, the authors attempted to identify a PHQ-9 cutoff that would match SCID-based prevalence, but heterogeneity was too high to generate consistently accurate estimates in individual studies for any PHQ-9 cutoff.

The Hospital Anxiety and Depression Scale (HADS) is a self-report screening questionnaire designed to be administered to non-psychiatric medical patients. It includes 14 items, with 7 assessing symptoms of depression (HADS-D) and 7 assessing symptoms of anxiety (HADS-A) over the past week. To avoid overlap with physical illness, the HADS-D does not include symptoms common to both physical and mental disorders, such as insomnia, loss of appetite, or fatigue. Cutoff thresholds of ≥8 and ≥11 on the HADS-D are traditionally used as standard cutoffs for identifying people who may have depression [7]. Although not designed for this purpose, the HADS-D is also frequently used to report depression prevalence in primary research studies. A review of recent studies listed in PubMed (2018–2019) identified 32 studies that reported “prevalence” of depression based on a HADS-D cutoff, with ≥8 and ≥11 used in 66% and 16% of the studies, respectively (see supplementary material eMethods 1 and eTable 1).

Although other screening tools and commonly used cutoffs have been shown to overestimate depression prevalence, it is not clear whether this would be the case with the HADS-D. A previous study that investigated prevalence of major depression among survivors of acute myocardial infarction found a prevalence of 20% (10,785 participants, 8 studies) using structured interviews, compared to 16% using a HADS-D cutoff of ≥8 (863 participants, 4 studies), and 7% using ≥11 (830 participants, 4 studies) [8]. This was a between-study comparison, however, and no included studies administered both the HADS-D and a validated diagnostic interview.

The objectives of the present study were to use an IPDMA approach to (1) compare pooled prevalence based on HADS-D cutoffs of ≥8 and ≥11 with major depression prevalence based on the SCID; and (2) use a prevalence-matching approach to determine if any cutoff threshold on the HADS-D matches prevalence based on the SCID with sufficiently low heterogeneity that it could be used to accurately measure depression prevalence in future studies.

Section snippets

Methods

This study used a subset of data collected for an IPDMA of the diagnostic accuracy of the HADS-D for screening to detect major depression. Detailed methods of the IPDMA were registered in PROSPERO (CRD42015016761), and a protocol was published [9]. The present analysis was not included in the original IPDMA protocol, which focused only on diagnostic accuracy. A protocol for the present study was published on the Open Science Framework prior to initiating the study (https://osf.io/n5a3e/).

Results

The initial search for the main IPDMA found 10,015 unique titles and abstracts for potential eligibility. Of these, we excluded 9584 studies after reviewing titles and abstracts and 238 studies after full-text review. There were 193 eligible studies using data from 133 unique samples from which 75 (56.4%) contributed individual participant data. Authors also contributed data from 8 unpublished studies, resulting in a total of 83 datasets. For our main analyses, we excluded 42 studies that used

Discussion

Previous research has demonstrated that there may be substantial differences between screening tools and diagnostic tools in estimating depression prevalence [[4], [5], [6]]. In the present study, we found that the most commonly used HADS-D cutoff threshold for reporting depression prevalence of ≥8 overestimated depression prevalence (24.5%) substantially compared to SCID major depression prevalence (11.6%). A HADS-D cutoff of ≥11 underestimated prevalence only slightly in aggregate compared to

Contributors

  • BLevis, PC, JPAI, SM, SBP, RCZ (DEPRESSD Steering Committee Members), MH, ZI, CGL, NDM, MT (DEPRESSD Knowledge Users), ABenedetti, and BDT (DEPRESSD Directors) were responsible for the conception, design and oversight of the main IPDMA project of which the present study is a part.

  • EB, DN, BLevis, JPAI, ABenedetti, and BDT were responsible for the conception and design of the present study.

  • JTB and LAK designed and conducted database searches to identify eligible studies.

  • ABeraldi, APBMB, GC, KC,

Roles of the funding source

This study was funded by the Canadian Institutes of Health Research (CIHR, KRS-144045 & PCG 155468). Ms. Neupane was supported by a G.R. Caverhill Fellowship from the Faculty of Medicine, McGill University. Drs. Levis and Wu were supported by Fonds de recherche du Québec - Santé (FRQS) Postdoctoral Training Fellowships. Mr. Bhandari was supported by a studentship from the Research Institute of the McGill University Health Centre. Ms. Rice was supported by a Vanier Canada Graduate Scholarship.

Declaration of Competing Interest

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf and declare that: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years with the following exceptions: (1) Dr. Ismail declares that he has received personal fees from Avanir, Janssen, Lundbeck, Otsuka, Sunovion, outside the submitted work. (2) Dr. Tonelli

Acknowledgements

We thank Dr. Linda Kwakkenbos for helping with translation and coding throughout the project.

References (14)

There are more references available in the full text version of this article.

Cited by (0)

View full text