Depression is a widespread mental disorder affecting 280 million individuals worldwide. It is also a prominent cause of disability and contributes to the global burden of disease.1 Experiencing pleasure is essential for an individual's well-being.Reference Berridge Kent and Kringelbach Morten2 However, individuals with mental disorders often have a diminished capacity to experience pleasure.Reference Trøstheim, Meir, Hansen, Paul and Kroll3 This inability to experience pleasure is referred to as anhedonia, a term that also describes an individual's lack of response to pleasant stimuli. Anhedonia has profound implications, potentially leading to decreased quality of life.Reference Vinckier, Gourion and Mouchabac4,Reference Ho and Sommers5 It is among the core symptoms of depression6 and is recognised as a transdiagnostic symptom across various mental disorders.Reference Guineau, Ikani, Rinck, Collard, van Eijndhoven and Tendolkar7 A large population-based study has shown that brain structure is related to phenotypes of and genetic risk factors for anhedonia; however, the study had limitations in terms of access to its target populations.Reference Zhu, Ward, Cullen, Lyall, Strawbridge and Lyall8 Therefore, the use of a self-report scale to assess anhedonia is advantageous because it enables researchers to easily reach their target populations. An example of such a scale is the Snaith–Hamilton Pleasure Scale (SHAPS), a 14-item self-report instrument that encompasses four domains of hedonic experience, namely: interest/pastimes, social interaction, sensory experience and food/drink. The scale was developed with consideration of culture, gender and age to minimise potential biases.Reference Snaith, Hamilton, Morley, Humayan, Hargreaves and Trigwell9 The SHAPS has been extensively used for assessing anhedonia in patients with major depressive disorder (MDD) and the general population across different countries and has been demonstrated to exhibit robust psychometric properties.Reference Zhang, Zhang, Fang, He, Fan and Luo10 The scale is also concise and easily comprehensible and has thus been translated into multiple languages, including Spanish, simplified Chinese, Italian and Malay.Reference Fresán and Berlanga11–Reference Ng, Chin, Yee, Loh, Sulaiman and Sherianne Sook Kuan14 Despite its widespread application, the SHAPS does not have a standard scoring method. The original scoring method involves the use of a dichotomous scoring system, with 0 indicating agreement and 1 indicating disagreement, and a total score ranging from 0 to 14. However, this scoring system may limit the ability of researchers to distinguish levels of severity for each item, to calculate relevant validity measures and to establish the correlations of SHAPS results with those of other instruments. To address these limitations, researchers have modified the scoring method by adopting a four-point Likert scale with anchors that range from ‘definitely agree’ to ‘definitely disagree,’ yielding a total score ranging from 14 to 56.Reference Snaith, Hamilton, Morley, Humayan, Hargreaves and Trigwell9,Reference Liu, Wang, Zhu, Li and Chan12,Reference Franken, Rassin and Muris15 Furthermore, SHAPS reference values have yet to be established because of the lack of a standard scoring method for the scale. In the original version of the SHAPS, individuals are considered to exhibit anhedonia if they disagree with more than two SHAPS items.Reference Snaith, Hamilton, Morley, Humayan, Hargreaves and Trigwell9 A systematic review and meta-analysis of 168 studies reported that the mean reference values for the SHAPS were 20.2 (s.d. = 2.1) and 33.1 (s.d. = 2.7) for nonclinical and clinical groups, respectively.Reference Trøstheim, Meir, Hansen, Paul and Kroll3 Nevertheless, the characteristics and severity of anhedonia may vary between clinical and nonclinical groups;Reference Trøstheim, Meir, Hansen, Paul and Kroll3 therefore, comparing SHAPS cut-off values between these two groups is imperative. Moreover, studies have shown that anhedonia is a multifaceted symptom that can be influenced by genetic, social and biological factors.Reference Ho and Sommers5,Reference Ren, Fabbri, Uher, Rietschel, Mors and Henigsberg16 Therefore, determining the predictors of anhedonia in both clinical and nonclinical groups is crucial for understanding its underlying mechanism and establishing appropriate intervention and policy measures.
As mentioned above, studies have confirmed a positive correlation between anhedonia and depressionReference Zhang, Zhang, Fang, He, Fan and Luo10,Reference Liu, Wang, Zhu, Li and Chan12 and have reported reference values for the SHAPS for both clinical and nonclinical groups.Reference Trøstheim, Meir, Hansen, Paul and Kroll3 Nevertheless, few of these studies established SHAPS cut-off values through receiver operating characteristic (ROC) curve analysis or used these values to compare predictors of anhedonia between clinical and nonclinical groups. The Patient Health Questionnaire-9 (PHQ-9) is a reliable, valid, brief and easily administered tool that has been validated for diagnosis of MDD.Reference Kroenke17–Reference Pettersson, Boström, Gustavsson and Ekselius19 Accordingly, in the present study, we used the PHQ-9 as a gold standard to determine cut-off values for the SHAPS and then used these values to compare predictors of anhedonia between clinical and nonclinical groups. On the basis of the findings of the aforementioned studies, we hypothesised that a significant correlation would exist between SHAPS and PHQ-9 scores in both clinical and nonclinical groups (hypothesis 1); that SHAPS cut-off values would differ between the clinical and nonclinical groups, with the clinical group having a higher cut-off values (hypothesis 2); and that the prevalence and predictors of high-risk depression, as determined using the SHAPS scores, would differ between the clinical and nonclinical groups categorised according to the optimal cut-off values (hypothesis 3).
Method
Research design and study participants
This study employed a cross-sectional and correlational research design. To derive SHAPS cut-off values for use in identifying and comparing predictors of anhedonia between clinical and nonclinical groups, we used convenience sampling to concurrently recruit 160 out-patients with MDD (i.e. the clinical group) from three hospitals and 412 university students (i.e. the nonclinical group) from two universities in northern Taiwan between August 2021 and May 2023. The inclusion criteria for the clinical group were as follows: (a) being aged 18–65 years and an out-patient with MDD; (b) receiving a clinical diagnosis of MDD, confirmed by experienced psychiatrists on the basis of ICD-10 codes (i.e. F32: depressive episode; F33: recurrent depressive disorder; F34: persistent affective disorder; F38: other affective disorder; F39: unspecified affective disorder); (c) being able to communicate in Mandarin Chinese; and (d) being willing to provide informed consent for participation in this study. The inclusion criteria for the nonclinical group were as follows: (a) being a university student; (b) being able to communicate in Mandarin Chinese; and (c) being willing to provide informed consent for participation in this study. Patients with any other psychiatric comorbidity or risk of suicide were excluded from the study. To determine the appropriate sample size for our study, we used G*Power 3.1,Reference Faul, Erdfelder, Buchner and Lang20 setting the medium effect size to 0.15, significance level to 5% and power to 80%; thus, we derived a total sample size of 118. In addition, we considered the widely accepted recommendation that correlational and comparative research requires a sample size of at least 30 participants in each group.Reference Gay, Mills and Airasian21 Thus, our final sample size ensured adequate statistical power for the study.
Measurements
Demographic characteristics
We collected information on the demographic characteristics of the clinical group, including gender, age, height, weight, educational level, geographical area, family status, socioeconomic status, working status, mental-health-related medical status and substance use status. For the nonclinical group, we collected information on demographic characteristics including gender, age, height, weight, geographical area, family status, socioeconomic status, working status, parenting style, available support resources, areas of concern, mental-health-related medical status and substance use status.
SHAPS
The SHAPS was applied to assess anhedonia in this study. The SHAPS is a self-report tool comprising 14 items designed to evaluate recent hedonic experiences across four domains: interest/pastimes, social interaction, sensory experience and food/drink.Reference Snaith, Hamilton, Morley, Humayan, Hargreaves and Trigwell9 As mentioned, various language versions of the SHAPS have been developed, including simplified and traditional Chinese versions. The simplified Chinese version of the SHAPS contains items that are rated on a four-point Likert scale with answers ranging from ‘definitely agree’ to ‘definitely disagree’; the total score of this scale ranges from 14 to 56. The test–retest reliability and Cronbach's α value of this scale have been reported to be 0.64 and 0.85, respectively.Reference Liu, Wang, Zhu, Li and Chan12 Moreover, the traditional Chinese version of the SHAPS is easy to read (N = 13, M = 8.85) and straightforward to answer (N = 13, M = 9.23), indicating favourable face validity. The test–retest reliability and Cronbach's α value derived for this scale were 0.87 and 0.91, respectively. Furthermore, the traditional Chinese version of the SHAPS has been reported to demonstrate positive correlations with PHQ-9 (r = 0.52, P < 0.001) and the Positive and Negative Suicide Ideation – Negative Suicidal Idea (PANSI-NSI) (r = 0.28, P < 0.01) as well as negative correlations with self-esteem (r = −0.51, P < 0.001) and the Positive and Negative Suicide Ideation – Positive Idea (PANSI-PI) (r = −0.54, P < 0.001); therefore, the Chinese version of the SHAPS can be considered to be reliable and valid.Reference Ho, Wu, Chen, Wang, Lee and Chung22
PHQ-9
PHQ-9 is a tool designed to detect the presence of depressive symptoms and measure the severity of depression experienced over the preceding 2 weeks. PHQ-9 comprises nine questions, each of which is rated on a four-point scale with end-points ranging from 0 (‘never’) to 3 (‘almost every day’). The total score ranges from 0 to 27, with a score of 5–9 indicating mild depression, 10–14 indicating moderate depression, 15–19 indicating moderately severe depression and ≥20 indicating severe depression.Reference Kroenke, Spitzer and Williams18,Reference Kroenke, Spitzer, Williams and Löwe23 The Chinese version of PHQ-9 has been reported to exhibit good reliability and validity in the Taiwanese population, with an internal consistency of 0.80 for adults and 0.77 for elderly people. In terms of concurrent validity, PHQ-9 exhibited a positive correlation with the Hamilton Rating Scale for Depression (r = 0.66, P < 0.001) and a negative correlation with the Quality of Life and Satisfaction Survey Volume (r = −0.53, P < 0.001).Reference Chen, Liu, Huang, Sun, Huang and Sung24,Reference Liu, Yeh, Huang, Sun, Tjung and Hwang25
Data collection
As the data collection period overlapped with the COVID-19 pandemic, data from both the clinical and nonclinical groups were collected either online or in person as convenient. For the clinical group, experienced psychiatrists identified potential participants during clinical consultations. Subsequently, research assistants reached out to these potential participants to provide a detailed explanation of the study's purpose and procedures. Those who expressed interest in participating in the study signed informed consent forms and were given the option to complete the research questionnaires either online or in person. For the nonclinical group, prospective students were approached either online or in person during scheduled class times. They were presented with an explanation of the study's purpose and procedures. Those who agreed to participate in the study signed informed consent forms and proceeded to complete the study questionnaires. After completing the questionnaires, all participants received NT$100 vouchers as a token of appreciation for their participation.
Statistical analysis
Data analysis was conducted using SPSS 24.0. The basic characteristics of the participants are presented as frequencies, percentages, means and s.d. values. Pearson correlation coefficients were derived to evaluate the correlations between the SHAPS and PHQ-9 scores in both the clinical and nonclinical groups. ROC curve analysis was conducted to determine the optimal cut-off values for the area under the ROC curve (AUC) and the sensitivity and specificity of the SHAPS in both the clinical and nonclinical groups. Independent-samples t-test, χ2-test and multiple logistic regression were used to determine the predictors and determinants of anhedonia on the basis of the SHAPS cut-off values. Furthermore, we conducted a sensitivity analysis to assess the reliability of the ROC curve analysis results in the clinical and nonclinical groups using two approaches. First, the data from both groups were combined and coded as 1 and 0 to replace PHQ-9 as the gold standard for ROC curve analysis. Second, to minimise the potential impact of variations in participant characteristics, SAS 9.4 was used to pair participants in the clinical group with those in the nonclinical group at a 1:2 ratio according to gender, age, height and weight; subsequently, ROC curve analysis was performed, and the results obtained with pairing were compared with those obtained without pairing. Statistical significance was set at 0.05 in this study.
Ethical considerations
This study protocol was approved by the Joint Institutional Review Board of Taipei Medical University, Taiwan (TMU-JIRB-N202011061) and the Institutional Review Board of National Yang Ming Chiao Tung University, Taiwan (YM110170E). The collected data were pseudonymised to ensure that no personally identifiable information was included. Only the researcher conducting this study had access to these data. Moreover, the data will be securely stored for a minimum of 7 years.
Results
Participant characteristics
The nonclinical group comprised 412 students, of whom 67% were women. On the basis of PHQ-9 scores, 174 (42.23%) of the students in the nonclinical group were classified as healthy, 139 (33.74%) were classified as having mild depression, 61 (14.81%) were classified as having moderate depression, 26 (6.31%) were classified as having moderately severe depression and 12 (2.91%) were classified as having severe depression. The mean age, height and weight of students in this group were 21.06 years (s.d. = 1.28; range, 18–25 years), 165.01 cm (s.d. = 7.99; range, 149–186 cm) and 58.52 kg (s.d. = 12.8; range, 36–113 kg), respectively. The clinical group comprised 160 patients, of whom 66% were women. On the basis of PHQ-9 scores, 25 (15.62%) of the patients in this group were classified as healthy, 31 (19.37%) were classified as having mild depression, 32 (20%) were classified as having moderate depression, 33 (20.63%) were classified as having moderately severe depression and 39 (24.38%) were classified as having severe depression. The mean age, height, and weight of patients in the clinical group were 39.49 years (s.d. = 13.91; range, 19–67 years), 163.8 cm (s.d. = 8.48; range, 146–185 cm) and 63.16 kg (s.d. = 15.23; range, 39–123 kg), respectively. We observed significant differences between the two groups in terms of depression levels, age and weight. The baseline characteristics of the participants are presented in Table 1.
Baseline characteristics of participants in the clinical and nonclinical groups (N = 572)
PHQ-9, Patient Health Questionnaire-9.
a. Fisher's exact test.
*** P < 0.001.
Correlations between SHAPS and PHQ-9 scores in clinical and nonclinical groups
Pearson correlation coefficients were computed to assess the linear relationship between SHAPS and PHQ-9 scores in both the clinical and nonclinical groups. The results revealed a positive correlation between the two instruments in both groups (r = 0.58, P < 0.001 and r = 0.41, P < 0.001, respectively; Table 2).
Pearson correlation analysis to assess the linear relationship between SHAPS and PHQ-9 scores in both the clinical and nonclinical groups (N = 572)
SHAPS, Snaith–Hamilton Pleasure Scale; PHQ-9, Patient Health Questionnaire-9.
*** P < 0.001.
Comparison of SHAPS cut-off values between clinical and nonclinical groups
We plotted ROC curves to derive SHAPS cut-off values for distinguishing various levels of depression severity, as determined using PHQ-9, in the clinical and nonclinical groups (Fig. 1). The results indicated that in the clinical group, the SHAPS could significantly differentiate between mild, moderate, moderately severe and severe depressive symptoms, with AUCs of 0.79, 0.76, 0.83 and 0.75, respectively (P < 0.001). Moreover, the optimal cut-off values corresponding to these levels of depression were determined using the Youden index,Reference Youden26 yielding cut-off values of 28.5, 28.5, 29.5 and 32.5 for mild, moderate, moderately severe and severe depressive symptoms, respectively (Table 3). The results also indicated that in the nonclinical group, the SHAPS could significantly distinguish between mild, moderate, moderately severe and severe depressive symptoms, with AUCs of 0.64, 0.72, 0.77 and 0.83, respectively (P < 0.001). The Youden indexReference Youden26 was again used to determine optimal cut-off values; these were 20.5, 21.28, 24.5 and 23.5 for mild, moderate, moderately severe and severe depressive symptoms, respectively (Table 3). The cut-off value associated with maximum sensitivity (78%) and specificity (76%) for the clinical group was 29.5, whereas that associated with maximum sensitivity (92%) and specificity (68%) for the nonclinical group was 23.5 (Table 3).
Receiver operating characteristic curves illustrating Snaith–Hamilton Pleasure Scale cut-off values for distinguishing between different severity levels of depression and clinical depression, as derived using Patient Health Questionnaire-9 scores and clinical diagnosis. (a) clinical group, (b) nonclinical group, (c) overall combined group.
Optimal cut-off values for Snaith–Hamilton Pleasure Scale for the clinical, nonclinical and combined overall groupsa
a. Optimal cut-off values based on the Youden index are presented in bold.
Sensitivity analysis
First, all participants in the clinical group received their diagnoses from experienced psychiatrists through structured clinical interviews, a widely accepted gold standard for psychiatric diagnoses.Reference First and Williams27 Accordingly, we conducted a sensitivity analysis to ascertain the reliability of our ROC curve analysis. For this analysis, we combined data from both the clinical and nonclinical groups and assigned them binary codes of 1 and 0 as substitutes for PHQ-9, which is the gold standard. Subsequently, we generated ROC curves to establish SHAPS cut-off values to effectively distinguish depression, as diagnosed clinically, in the overall combined group (Fig. 1). In the combined group, SHAPS demonstrated significant discriminatory ability for clinical depression, with a corresponding AUC of 0.8 (P < 0.001). Moreover, the optimal cut-off values for clinical depression were determined using the Youden index.Reference Youden26 The optimal cut-off values for clinical depression were identified as 23.5 with a sensitivity of 81% and specificity of 66%, and 24.5 with a sensitivity of 76% and specificity of 70%, in the overall combined group (Table 3). These results were consistent with the optimal cut-off values determined for the nonclinical group.
Second, we paired the participants in the clinical group with those in the nonclinical group at a 1:2 ratio and then conducted ROC curve analysis. The characteristics of the participants in the clinical and nonclinical groups did not differ significantly when the participants were paired according to gender (P = 1), age (P = 0.3), height (P = 0.61) or and weight (P = 0.13). Moreover, the ROC curve analysis results obtained for these paired groups were consistent with those obtained for the original groups.
Prevalence and predictors of high-risk depression based on optimal SHAPS cut-off values in both clinical and nonclinical groups
We used the optimal SHAPS cut-off values to estimate the prevalence of high-risk depression in the two groups. The estimated prevalence of high-risk depression was 48.1% in the clinical group and 34% in the nonclinical group. Furthermore, we showed by chi-squared test that high-risk depression was significantly correlated with self-awareness of depression (P = 0.002) and self-harm history (P = 0.03) in the clinical group. An independent-samples t-test also indicated that high-risk depression was significantly correlated with lower academic performance (P < 0.001) and academic satisfaction (P < 0.001) scores in the nonclinical group. In addition, chi-squared test results demonstrated that in the nonclinical group, high-risk depression was significantly correlated with skipping classes (P < 0.001); father's parenting attitude (P = 0.01); mother's parenting attitude (P = 0.02); help-seeking behaviours, including seeking help from relatives (P < 0.001) and friends (P = 0.006); receiving psychological counselling (P = 0.012); and seeing a psychiatrist (P = 0.002; Table 4).
Predictors of high-risk depression based on optimal cut-off values for Snaith–Hamilton Pleasure Scale
We finally conducted multiple logistic regression analyses, including factors such as self-awareness of depression and self-harm history for the clinical group and factors such as academic performance, academic satisfaction, skipping classes, parenting attitude, help-seeking behaviours, receiving psychological counselling and seeing a psychiatrist for the nonclinical group. These factors were included because they differed significantly between participants at high and low risk of depression in both groups (Table 4).
In the clinical group, compared with the participants with a low risk of depression, those who believed that they were not depressed (odds ratio = 0.08; 95% CI 0.01–0.64) had a lower probability of having a high risk of depression. In the nonclinical group, compared with participants with a low risk of depression, those who had lower academic satisfaction (odds ratio = 1.55; 95% CI 1.09–2.21) and were not willing to seek help from relatives (odds ratio = 1.80; 95% CI 1.09–2.96) or friends (odds ratio = 1.79; 95% CI 1.09–2.96) had a higher probability of having a higher risk of depression. Moreover, participants whose fathers’ parenting attitudes were characterised by greater care and protection (odds ratio = 0.4; 95% CI 0.18–0.89) and those who were not skipping classes (odds ratio = 0.48; 95% CI 0.25–0.9) had a lower probability of having a high risk of depression (Table 5).
Determinants of high-risk depression based on the optimal cut-off values for the Snaith–Hamilton Pleasure Scale
Ref., reference group (low-risk depression group).
* P < 0.05.
Discussion
To the best of our knowledge, this is the first study to establish SHAPS cut-off values for both clinical and nonclinical populations using ROC curve analysis. We calculated Pearson correlation coefficients to assess the correlation between SHAPS and PHQ-9 scores and used ROC curve analysis to determine optimal SHAPS cut-off values for both clinical and nonclinical groups. Our findings demonstrate a significant correlation between SHAPS and PHQ-9 scores. Moreover, the optimal SHAPS cut-off values derived in this study could effectively distinguish between individuals at risk of depression in both the clinical and nonclinical groups.
We observed a significant correlation between SHAPS and PHQ-9 scores in both the clinical and nonclinical groups; thus, hypothesis 1 is supported. This finding is also consistent with those of previous studies.Reference Zhang, Zhang, Fang, He, Fan and Luo10,Reference Liu, Wang, Zhu, Li and Chan12 The correlation coefficient for the association between SHAPS and PHQ-9 scores in the clinical group was higher (moderate positive) than that in the nonclinical group (moderate positive). Furthermore, we noted that the optimal SHAPS cut-off values differed between the clinical and nonclinical groups, with the clinical group having higher cut-off values; thus, hypothesis 2 is supported. This finding is consistent with those of previous research.Reference Trøstheim, Meir, Hansen, Paul and Kroll3 The differences in cut-off values may be attributed to variations in the severity of anhedonia and baseline depression levels between the clinical and nonclinical groups. Studies have demonstrated that anhedonia is more prevalent in patients with MDD (up to 70%)Reference Shankman, Katz, DeLizza, Sarapas, Gorka, Campbell and Ritsner28 than in the general population (approximately 20%).Reference Werbeloff, Dohrenwend, Yoffe, van Os, Davidson and Weiser29 PHQ-9 is widely used to evaluate the severity of depression, whereas the SHAPS is used to assess anhedonia, a key symptom of depression.Reference Liu, Wang, Zhu, Li and Chan12,Reference Chen, Liu, Huang, Sun, Huang and Sung24,Reference Liu, Yeh, Huang, Sun, Tjung and Hwang25 Therefore, clinical populations may have higher SHAPS scores and a stronger correlation with anhedonia when compared with nonclinical populations; this is consistent with our findings (Table 1) and those of a previous study.Reference Zhang, Zhang, Fang, He, Fan and Luo10
The optimal SHAPS cut-off values derived in this study could help to determine the prevalence and predictors of high-risk depression in clinical and nonclinical groups; thus, hypothesis 3 is supported by our results. The optimal SHAPS cut-off was 23.5 in the nonclinical group; on the basis of this cut-off value, we determined that the prevalence of high-risk depression was 34% in this group. This is consistent with the findings of the Global Point Prevalence Survey for elevated depressive symptoms and with the results of a population-based cohort study.Reference Ho, Chiou, Molloy, Lin, Chang and Chang30,Reference Shorey, Ng and Wong31 Accordingly, the optimal SHAPS cut-off (23.5) derived for the nonclinical group has the potential to distinguish high-risk individuals from a target population. Furthermore, the optimal SHAPS cut-off was 29.5 in the clinical group; on the basis of this cut-off value, we determined that the prevalence of high-risk depression was 48.1% in this group. This finding is consistent with those of previous studies, which have demonstrated that nearly 50% of patients with MDD did not experience major improvements after routine treatment, and that the recurrence rate of MDD after the first episode was approximately 50%.Reference Bains and Abdijadid32,Reference Bear, Edbrooke-Childs, Norton, Krause and Wolpert33 Hence, the optimal SHAPS cut-off (29.5) derived in this study for the clinical group could serve as an early indicator of poor treatment response or recurrence risk in individuals with MDD. In addition, our logistic regression results show that participants who perceived that they were not depressed exhibited more favourable outcomes compared with those who perceived that they were depressed. This finding raises questions about the role of patient belief and self-awareness of depression. Therefore, we subsequently analysed the correlation between self-awareness of depression and emotional conditions on the basis of the optimal SHAPS cut-off values derived for the two groups. The results revealed that 70% of participants who had scores that exceeded the optimal cut-off values believed that they were depressed, whereas 93% of participants who did not meet the criteria believed that they were not depressed. The significant correlation between self-awareness of depression and emotional condition suggests that patient insight plays a crucial part in the patient's recognition and acknowledgement of their emotional state. Therefore, we confirmed that a substantial proportion of the participants with depression had a reasonable awareness of their condition. This finding aligns with those of previous research, which has reported that participants with MDD exhibited better insight into their conditions, and that 36.8% of the participants had impaired insight into their conditions.Reference Weiler, Fleisher and McArthur-Campbell34,Reference Yen, Chen, Lee, Tang, Ko and Yen35 The results thus indicate that the optimal cut-off value established for the clinical group could be a valuable tool to monitor the risk of recurrence in patients with poor insight into their depression. Furthermore, our logistic regression model identified risk factors for depression in the nonclinical group, including lower academic satisfaction and poor help-seeking behaviours. These findings are consistent with those of previous research, which demonstrated a significant correlation between lower academic satisfaction and depression (β = −0.26, P < 0.001);Reference Tran, Franzen, Jermann, Rudaz, Bondolfi and Ghisletta36 they are also consistent with the results of a large-scale national survey that revealed a significant correlation between decreased help-seeking behaviours and depressive symptoms.Reference Oh, Hong, Oh and Kim37 Moreover, the model showed protective factors against depression in the nonclinical group, including father's parenting attitude and good attendance at school. These results are also consistent with those of previous studies, which have emphasised the importance of fathers’ involvement and support, as well as regular school attendance, in reducing depressive symptoms.Reference Askeland, Bøe, Lundervold, Stormark and Hysing38,Reference Opondo, Redshaw and Quigley39 Our findings underscore the significance of the father's parenting attitude in mitigating depression in the nonclinical group. The absence of significance regarding the mother's parenting attitude in the final model may be attributed to cultural factors prevalent in Asian parenting. Research has demonstrated that the authoritarian parenting style, characterised by coldness, lack of support and stringent control, is associated with lower resilience and a higher risk of depression.Reference Zhong, Wu, Nie, Xia, Li and Lei40 Furthermore, traditional Chinese fathers typically wield more power and authority over their children than mothers.Reference Lamb41 Consequently, the observed importance of high-care and high-protection parenting attitudes of fathers in this study suggests that such attitudes, in contrast to the authoritarian style, may act as a significant protective factor against depression in nonclinical populations.
Our findings reveal a significant correlation between SHAPS and PHQ-9 scores. In addition, we derived optimal SHAPS cut-off values for the clinical and nonclinical groups and then used these values to identify risk factors for high-risk depression in both groups. Notably, these cut-off values and risk factors differed between the clinical and nonclinical groups; therefore, developing strategies and interventions tailored to either group could help to enhance the effectiveness of primary and secondary preventive measures. For example, routine health education programmes could be implemented across diverse populations to help educate them on how to use the SHAPS for self-reflection and to document emotional changes, which could ultimately facilitate the implementation of primary preventive measures. Moreover, the study highlights the potential of SHAPS and its optimal cut-off values as early indicators of depression risk, especially in clinical settings. Identifying individuals at high risk of recurrence or poor treatment response could enable the provision of timely interventions and support. Intervention programmes should also consider both group-level patterns (optimal SHAPS cut-off values) and individual-level changes in SHAPS scores to maximise their effectiveness. Furthermore, the risk factors identified in this study could serve as event markers for emotional support and crisis intervention. Monitoring and intervention programmes should be tailored to an individual's characteristics. For example, individuals at high risk of recurrence may benefit from more intensive support and relapse-prevention strategies, whereas those in nonclinical settings may require educational interventions.
This study had several strengths. First, we concurrently recruited participants from two schools and three hospitals across different districts in Taiwan to explore SHAPS cut-offs and the correlations between SHAPS and PHQ-9 scores in both clinical and nonclinical groups, which increased the external validity of our findings. Second, the inclusion of data from both clinical and nonclinical groups allowed us to conduct clinical discrimination, derive optimal SHAPS cut-off values and use these values to identify predictors of anhedonia. Third, in contrast to most studies in the literature, which have focused on developing and validating SHAPS items without establishing or comparing SHAPS cut-off values within clinical and nonclinical groups, the present study employed PHQ-9 as a gold standard to determine the optimal SHAPS cut-off values and subsequently identify relevant risk factors within the clinical and nonclinical groups. In addition, we employed clinical and nonclinical categories (coded as 1 and 0) as the gold standard instead of PHQ-9 to validate the feasibility of the derived cut-off values. The results not only confirmed the feasibility of the derived cut-off values but also demonstrated the validity of PHQ-9 as a gold standard.
This study also had some limitations that should be acknowledged. First, we adopted a cross-sectional research design. Therefore, we could not fully establish causal relationships between risk factors and SHAPS scores. Second, although we conducted multiple sensitivity analyses to validate our results, the effects of participant characteristics must be considered. Accordingly, our findings should be interpreted with caution, and longitudinal studies with multicentre sampling should be conducted in the future. Finally, the SHAPS is not a diagnostic instrument but a screening tool for identifying individuals at high risk of depression or recurrence. Therefore, individuals whose SHAPS scores reach an established cut-off value should undergo further assessment and evaluation and should receive early interventions, if necessary.
This study confirms that there is a significant correlation between SHAPS and PHQ-9 scores, highlighting the utility of the SHAPS as a tool for assessing anhedonia in individuals. In addition, the optimal SHAPS cut-off values and the predictors of anhedonia determined using these values differed between the clinical and nonclinical groups. These findings demonstrate that such optimal cut-off values and predictors can be useful in the screening of individuals at high risk of depression or recurrence. Our findings study also emphasise the need for comprehensive, individualised programmes using smart devices to screen for anhedonia. These programmes could use SHAPS scores and relevant predictors to identify at-risk individuals and provide early interventions. Furthermore, future research could benefit from longitudinal follow-up studies that monitor SHAPS scores and related predictors over time. This approach would enable a deeper understanding of the dynamic nature of anhedonia and provide opportunities for timely interventions.
Data availability
The data that support the findings of this study are available from the corresponding author, H.-J.C., upon reasonable request.
Acknowledgements
We thank the students who participated in the study.
Author contributions
Study design: Y.-C.H., H.-J.C. Data collection: Y.-C.H, Y.-S.W., Y.-C.L. Data analysis: Y.-C.H. Revisions for important intellectual content: S.S.-F.G., C.-H.C., J.-K.W, H.-C.L., K.-H.C., Y.-H.C., K.-K.G, M.-L.L., H.-J.C., P.-C.C.
Funding
This study was supported by the National Science and Technology Council in Taiwan.
Declaration of interest
None.
Open access
eLetters
No eLetters have been published for this article.