How symptoms of simple acute infections affect the SSS-8 and SSD-12 as screening instruments for somatic symptom disorder in the primary care setting

Measurement Properties of the Patient Health Questionnaire-15 and Somatic Symptom Scale-8: A Systematic Review and Meta-Analysis

Importance

The subjective experience of somatic symptoms is a key concern throughout the health care system. Valid and clinically useful instruments are needed.

Objective

To evaluate the measurement properties of 2 widespread patient-reported outcomes: the Patient Health Questionnaire-15 (PHQ-15) and Somatic Symptom Scale-8 (SSS-8).

Data Sources

Medline, PsycINFO, and Web of Science were last searched February 1, 2024.

Study Selection

English-language studies reporting estimates pertaining to factor analysis, taxometric analysis, internal consistency, construct validity, mean scores in relevant groups, cutoffs, areas under the receiver operating characteristic curves (AUROCs), minimal clinically important difference, test-retest reliability, or sensitivity to change.

Data Extraction and Synthesis

Search hits were reviewed by independent raters. Cronbach α, Pearson r, means, and between-group effect sizes indicative of sensitivity to change were pooled in random-effects meta-analysis. Study quality was assessed using 3 instruments. Reporting followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 reporting guideline.

Main Outcomes and Measures

Comprehensive overview of evidence pertaining to the measurement properties of the PHQ-15 and SSS-8.

Results

A total of 305 studies with 361 243 participants were included. Most concerned routine care (178 studies) and the general population (27 studies). In factor analyses, both scales reflected a combination of domain-specific factors (cardiopulmonary, fatigue, gastrointestinal, pain) and a general symptom burden factor. The pooled PHQ-15 α was 0.81 (95% CI, 0.80-0.82), but with low item-total correlations for items concerning menstrual problems, fainting spells, and sexual problems (item-total correlations <0.40), and the SSS-8 α was 0.80 (0.77-0.83). Pooled correlations with other measures of somatic symptom burden were 0.71 (95% CI, 0.64-0.78) for the PHQ-15 and 0.82 (95% CI, 0.72-0.92) for the SSS-8. Reported AUROCs for identification of somatoform disorders ranged from 0.63 (95% CI, 0.50-0.76) to 0.79 (95% CI, 0.73-0.85) for the PHQ-15 and from 0.71 (95% CI, 0.66-0.77) to 0.73 (95% CI, 0.69-0.76) for the SSS-8. The minimal clinically important difference on both scales was 3 points. Test-retest reliability could not be pooled and was inconsistent for the PHQ-15 (PHQ-15: r = 0.65-0.93; ICC, 0.87; SSS-8: r = 0.996, ICC = 0.89). The PHQ-15 showed tentative sensitivity to change (g = 0.32; 95% CI, 0.08-0.56), but data for the SSS-8 were lacking.

Conclusions and Relevance

In this systematic review and meta-analysis, findings supported use of the PHQ-15 and SSS-8 for the assessment of symptom burden, but users should be aware of the complex, multifactorial structures of these scales. More evidence is needed concerning longitudinal measurement properties.

Lessons learned from applying established cut-off values of questionnaires to detect somatic symptom disorders in primary care: a cross-sectional study

Introduction

Based on two diagnostic accuracy studies in high-prevalence settings, two distinctly different combinations of cut-off values have been recommended to identify persons at risk for somatic symptom disorder (SSD) with the combination of the Patient-Health Questionnaire-15 (PHQ-15) and the Somatic Symptom Disorder-B Criteria Scale (SSD-12). We investigated whether the reported sensitivity and specificity of both recommended cut-off combinations are transferable to primary care.

Methods

In a cross-sectional study, 420 unselected adult primary care patients completed PHQ-15 and SSD-12. Patients scoring ≥9 and ≥ 23 (recommended cut-off combination #1) or ≥ 8 and ≥ 13 (recommended cut-off combination #2) were considered test-positive for SSD, respectively. To assess the validity of the reported sensitivity and specificity in different low- to high-prevalence settings, we compared correspondingly expected proportions of test positives to the proportion observed in our sample.

Results

Based on combination #1, 38 participants (9%) were found to be test positive, far fewer than expected, based on the reported values for sensitivity and specificity (expected minimum frequency 30% with a true prevalence ≥1%). This can only be explained by a lower sensitivity and higher specificity in primary care. For combination #2, 98 participants (23%) were test positive, a finding consistent with a true prevalence of SSD of 15% or lower.

Discussion

Our analyzes strongly suggest that the sensitivity and specificity estimates reported for combination #1 are not applicable to unselected primary care patients and that the cut-off for the SSD (≥23) is too strict. Cut-off combination #2 seems more applicable but still needs to be tested in studies that compare screening findings by questionnaires with validated diagnostic interviews as reference standards in primary care populations.