Health anxiety is defined by cognitive-behavioral researchers as worry about health ranging from mild concern to excessive preoccupation (Ferguson, 2009; Longley et al., 2010). Excessive and persistent health anxiety is a common and disabling condition that can result in substantial suffering, difficult doctor-patient relationships and high health care costs (Robbins and Kirmayer, 1996; Barsky et al., 2001; Fink et al., 2010; Sunderland et al., 2013; Bobevski et al., 2016). Moreover, health anxiety can become chronic and incapacitating, with a majority (>60%) of cases still showing symptoms after several years of follow-up (Sadock, 2017). Health anxiety often remains unrecognized (Gureje et al., 1997; Conradt et al., 2006); thus, early detection is crucial for reducing such serious impairment. Screening for health anxiety among primary care patients has been proposed (Fink et al., 1999).

Health anxiety has often been referred to as hypochondriasis (American Psychiatric Association., 2000). However, the DSM-IV definition of hypochondriasis has received criticism because the criteria were too narrow to be applied in clinical practice (Fink et al., 2004). Therefore, the DSM-V replaced hypochondriasis, among others, with illness anxiety disorder (IAD) and somatic symptom disorder (SSD) (American Psychiatric Association, 2013), and the classifications of IAD and SSD have proven to be more reliable in detecting health anxiety than hypochondriasis (Newby et al., 2017). The difference between the two disorders lies in the severity of somatic symptoms. SSD diagnosis requires distressing and disabling somatic symptoms in criterion A, and related thoughts, feelings, and behaviors in criterion B. IAD is characterized by non-existent or relatively minor somatic symptoms, excessive fears of illness, and high levels of health anxiety (Sadock, 2017). Bailer et al. (2016) found no difference in health anxiety severity, other hypochondriacial characteristics, illness behavior, somatic symptom attributions, and physical concerns between IAD patients and SSD patients, whereas Newby et al. (2017) proposed that SSD patients show more severe health anxiety. Nevertheless, these studies have suggested that health anxiety is an important feature of IAD and SSD.

The Whiteley Index (WI) is a “classic” scale of hypochondria, which was first developed by Pilowsky in the 1960s. Based on the principle diagnosis criteria of hypochondriasis in DSM IV, the scale consists of 14 items in 3 dimensions: disease phobia, somatic preoccupation and disease conviction (Pilowsky, 1967). Subsequent study has indicated a wide variation of factor models when generalized to different populations (Speckens et al., 1996). Hence, various of derivative of WI were developed, among which the most widely applied version is the 7-item WI (WI-7) developed by Fink et al. via latent structure analysis (Fink et al., 1999; Tu et al., 2016; Laferton et al., 2017). Due to the highly overlapped diagnosis criteria of hypochondria and other somatoform disorders, a more specific manifestation was investigated to disentangle hypochondria from somatoform disorder. In an interview study involving 701 participants, it was found that patients with health anxiety generally had the symptom “rumination,” which is not common in patients with somatoform disorder (Fink et al., 2004). Thus, the new item on obsessive rumination, “Recurring thoughts about having a disease that are difficult to get rid of,” was added to the original WI-7 (Carstensen et al., 2020). The updated 8-item (WI-8) may be more helpful in screening for health anxiety.

The aim of the present study was to examine the psychometric properties of an updated WI-8 version to better identify health anxiety in outpatients in general hospitals in China.

WI-8 was first developed by Fink et al. (2004) and Carstensen et al. (2020), and our study is the first to perform a psychometric evaluation of the Chinese version of the WI-8 as a health anxiety screening scale. Participants were outpatients in general hospitals, predominantly city dwellers, well-educated and not living alone. The results revealed good reliability, validity and discriminant ability, suggesting that the WI-8 could be used as an effective tool for screening health anxiety.

Cronbach's alpha of the WI-8 (0.937) was higher than that of all the previous versions (between 0.68 and 0.836) (Pilowsky, 1967; Welch et al., 2009; Lee et al., 2011; Hedman et al., 2015; Fergus et al., 2018). Moreover, Cronbach's alpha decreased after deleting the 8th item, indicating that the additional item 8 is not redundant, but increased the consistency of the scale. Various studies have reported the factor structure of WI and proposed unitary, 2-factor, and 3-factor models consisting of 6–14 items, indicating that the internal structure is unstable (Pilowsky, 1967; Welch et al., 2009). Therefore, the entire sample was randomly split into two parts: one group (n = 348) explores the WI-8 factor structure by EFA, the other group (n = 348) compares the goodness of fit of different factor structures by CFA. EFA indicated a single factor model for WI-8, suggesting that the newly added item did not alter the one-factor structure of the original WI-7. Previous studies have also proposed that the original three factors of WI are highly related, and the one-factor model may be more suitable (Speckens et al., 1996). While Fink et al. (2004) pointed out that two isolated factors were more consistent with clinical characteristics, and they proposed a two-factor model including Illness Conviction subscale and Illness Worrying subscale. However, we compared the fitting index of the two models and found that the two-factor model was not significantly better than the single-factor model. Furthermore, the AVE and CR indicated good convergent validity. It is possible that participants tended to rate all questions on the questionnaire consistently, resulting in a single structure.

The WI-8 had moderate correlations with the GAD-7, PHQ-9, and PHQ-15, whereas in a study involving 200 patients and 240 healthy graduate students, the WI-7 had lower correlations with the GAD-7 (0.35), PHQ-9 (0.30), and PHQ-15 (0.33) (Güleç et al., 2012), perhaps because these psychological factors were more closely linked in patients than in healthy people. Furthermore, the WI-8 had the weakest correlation with the PHQ-15 and the strongest correlation with the SSD-12. These findings are consistent with the IAD characteristics of light somatic symptoms and excessive health-related thoughts, and also suggest that health anxiety is an important feature of SSD.

The WI-8 showed excellent discriminatory ability to distinguish patients with health anxiety, which is consistent with WI-6 (AUC = 0.83) (Fergus et al., 2019), and superior to WI-7 (AUC = 0.660) (Tu et al., 2016). The results indicated that the WI-8 offers a better way to identify persons with health anxiety, likely because of the additional item emphasizing recurring thoughts.

This study has several limitations. The stability of the WI-8 over time could not be evaluated due to the cross-sectional study design. Additionally, the self-report scale could have been influenced by the educational background of the participants. Moreover, too few IAD patients could be analyzed as a subgroup, which affected the testing of the discriminant ability of the WI-8. Therefore, we grouped IAD patients and SSD patients into one group based on the similarity of their health anxiety characteristics. A lower proportion of IAD patients may be due to their mild physical symptoms and less hospital visits. Hence, future research with appropriate sample is planned to confirm and extend our findings in patients with general anxiety.

The latest version of the 8-item WI has satisfactory reliability, validity and discriminant ability in general hospital outpatients in China. Therefore, we introduce the WI-8 as an assessment tool for screening health anxiety.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the ethics committees of Peking Union Medical College Hospital (PUMCH) and the University Medical Centre, Freiburg, Germany (Protocol Number: S-K276). The patients/participants provided their written informed consent to participate in this study.

HW, KF, PF, JW, LZ, and WLi contributed conception and design of the study. YZ, HC, XM, JR, and WLu contributed to data collection. YC and RL performed the statistical analysis. YC wrote the first draft of the manuscript. HW, KF, RL, and A-KT wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted vision.

JR was employed by General Hospital of Jincheng Anthracite Coal Mining Group Co. Ltd., Jincheng, China. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.