SUMMARY:

Depression is a significant global contributor to disabilities, affecting individuals across various educational and socioeconomic levels. This study aimed to assess the prevalence and severity of depressive symptoms in the US population during the COVID-19 pandemic using the PHQ-8 questionnaire integrated into the Goodville mobile game. Data was collected from 72,123 US users of the app between April 2021 and May 2022. The study utilized a design that integrated the PHQ-8 questionnaire into the Goodville app's emotional well-being module, allowing users to assess their depressive symptoms. Validity of responses was assessed using the Rasch model, resulting in 54,856 valid response patterns for analysis. The findings revealed a high rate of depression, with 34.7% of users classified as clinically depressed. Low energy, sleep problems, and loss of interest were identified as the most common symptoms. Comparisons with previous studies showed consistency in both overall depression prevalence and the frequency distribution of symptoms. The configuration of symptoms remained largely unchanged, with somatic and motivational symptoms predominating. However, the severity of these symptoms significantly increased during the pandemic. The study demonstrated the validity of integrating assessment tools into mobile applications for population mental health data collection, providing scientifically informed insights into mental health and emotional well-being in broad populations. The findings contribute to understanding the impact of the COVID-19 pandemic on depression rates and symptomatology, facilitating targeted interventions and support for individuals experiencing depression.‍

Key words: depression prevalence, US population, mobile assessment, Goodville

There is no doubt that depression is a significant contributor to disabilities globally, leading to decreased functioning across various educational and socioeconomic levels. According to the World Health Organization (WHO), depression affects 3.8% of the world's population, including 5.0% of adults and 5.7% of adults older than 60. However, a study analyzing data from 30 countries (Grønli et al., 2022) reported an average depression prevalence of 12.9% across all studies, highlighting substantial heterogeneity between them. Diagnosing depression involves assessing the severity of symptoms, and understanding the distribution of these symptoms has garnered significant interest (Tomitaka et al., 2018). Subclinical depression symptoms persisting over time increase the risk of developing clinical depression. Even mild depressive symptoms can have adverse effects on daily functioning and quality of life. Numerous studies have utilized depression screening scales to investigate depressive symptoms in the general population. Commonly used scales include the Center for Epidemiologic Studies Depression Scale (CES-D), the 6-item Kessler Screening Scale for Psychological Distress (K6), and the Patient Health Questionnaire-9 (PHQ-9) (Astutik et al., 2021; Kessler et al., 2002; Tomitaka et al., 2018). While most studies focus on determining the severity of depression based on total scores and establishing optimal cutoff points, little attention has been given to the distributional patterns of depression symptoms assessed using these scales. In recent years, only a few studies have examined the frequency of depressive symptoms among the general population. A study by Tomitaka et al. (Tomitaka et al., 2018) found that the prevalence of depression symptoms in the US population between 2013 and 2014 ranged from 3% to 51%, depending on severity and type. The most common symptoms were low energy (51%) and sleep problems (37%), while the least common were suicidal ideation (3%) and psychomotor agitation (11%) in the US population (Tomitaka et al., 2018). According to Grønli et al.'s study (Grønli et al., 2022) on depression symptoms in the Norwegian population in 2015-2016, only 6.9% of respondents reported clinically significant depression symptoms. The authors attribute this low frequency of depressive symptoms to their use of a cut-off total score from the Hospital Anxiety and Depression Scale, rather than assessing individual symptoms. Over the past two years, the COVID-19 pandemic has significantly impacted depression symptoms in the general population. Research studies indicate that depression symptoms among US residents increased more than threefold during COVID-19 (Ettman et al., 2020). It is important to note that there is limited research on the prevalence of depressive symptoms during different stages of the COVID-19 pandemic.

In recent years, there has been a noticeable trend in collecting mental health data through mobile applications. Mobile apps offer integrated self-assessment questionnaires that can evaluate various mental health issues. The accessibility of mobile devices allows for rapid data collection related to mental health. The Goodville app, for example, incorporates the PHQ-8 questionnaire in its emotional well-being module, enabling users to assess their depressive symptoms. Over the past two years, a substantial amount of data has been gathered on the prevalence of depressive symptoms among US users of the Goodville app. Comparing these findings with existing studies will provide valuable insights into the reliability of the collected data. Additionally, a comparative analysis will investigate changes in the prevalence of depressive symptoms in the US during the COVID-19 pandemic compared to the pre-COVID period.

The objective of this study was to evaluate the prevalence of depression symptoms among players of the Goodville game from the US population during the period of COVID-19 (2021-2022) using PHQ-8 for mobile data collection.

The Patient Health Questionnaire-8 (PHQ-8) (Dhingra et al., 2011; Kroenke et al., 2009) is an adapted version of the PHQ-9, consisting of eight items. It is a self-report questionnaire designed to match symptoms with diagnostic criteria for identifying individuals with depression. The PHQ-8 excludes item 9 from the PHQ-9 to avoid potential inaccuracies in assessing suicide risk. As part of the emotional well-being module in the Goodville app, the PHQ-8 is integrated as a psychometric tool to monitor depression symptoms among Goodville users worldwide. Goodville players are asked to indicate the frequency of experiencing eight possible problems or symptoms (e.g., feeling down, depressed, or hopeless in the past week, feeling tired or lacking energy, feeling negative about oneself or expecting failure). Scale scores are calculated by summing up each item, with response options ranging from 0 (never) to 3 (almost every day).

Data collection took place between April 2021 and May 2022. A total of 72,123 respondents from the United States provided responses to the PHQ-8 items during this period. Prior to analyzing depressive symptoms, the validity of player responses was assessed using the Rasch model (Boone, 2016; Boone et al., 2014). Rasch person fit statistics were calculated for each response pattern, and response patterns with standardized mean-square Z-statistic values within the range of -1.9 to 1.9 were considered valid (Green & Frantom, 2002; B. D. Wright & Masters, 1990; B. Wright & Linacre, 1980). Response patterns not meeting this criterion were excluded from further analysis, resulting in 58,986 valid response patterns.

Statistical comparisons were conducted using Friedman's test for paired samples, followed by multiple pairwise comparisons using Nemenyi's method. The distribution of depression ranges and depressive symptoms was also analyzed visually.

Figure 1 presents the distribution of depression rates among a sample of US users based on the PHQ-8 interpretation criteria.

The data reveals that 44% of the players did not exhibit noticeable signs of depression. Mild depression was reported by slightly more than one-fifth of the respondents, while another one-fifth experienced moderate depression. The response rate for moderate to severe depression was 12%, and severe depression was reported by 3% of the participants.

A statistically significant difference was observed in depression symptoms according to Friedman's test (Q = 44808.583, p < 0.0001. The statistical analysis using Nemenyi's procedure (Table 1) revealed significant differences between the symptoms. Low energy (A) and sleep problems (B) showed the highest mean ranks, followed by loss of interest (C), indicating their greater prominence in the sample. Depressive mood (D) and appetite change (D, E) were also statistically significant but obtained slightly lower mean ranks compared to the aforementioned symptoms. Low self-esteem (E), difficulty concentrating (F), and slowness/agitation (G) exhibited progressively lower mean ranks, indicating a lesser degree of statistical significance and lower prominence among the assessed depressive symptoms.

According to Table 1, low energy, sleep problems, and lack of interest were the most pronounced depressive symptoms. These symptoms obtained the highest sum of ranks and mean of ranks among all the evaluated symptoms. On the other hand, psychomotor disturbances and concentration problems were the least pronounced symptoms, as indicated by their lower mean ranks.

The histogram presented in Figure 2 provides an overview of the distribution of depressive symptoms among Goodville users in the US population. Analysis of the histogram indicates that the prevalence of depression symptoms in the US population between 2021 and 2022 varied from 31% to 70%, depending on the severity and type of symptoms.

Of particular importance is the considerable prevalence of depression symptoms that occur more than half of the days or every day. This observation draws attention to the significant impact and persistence of these symptoms. Notably, three symptoms exhibit the highest prevalence: low energy levels (36%), sleep problems (36%), and loss of interest (26%).

In this study, we aimed to assess the prevalence and severity of depressive symptoms in the US population using the PHQ-8 questionnaire embedded in the Goodville mobile game. The majority of app users reported experiencing depressive symptoms of varying severity. Most respondents exhibiting signs of depression fell into the mild to moderate range. We observed a statistically significant difference in severity among the eight symptoms of depression measured by the PHQ-8. Notably, somatic symptoms such as low energy (70%) and sleep problems (64%) were the most prominent. It is worth mentioning that a fifth of respondents reported severe levels of these symptoms (Figure 2).

Based on the PHQ-8 scores (≥10) for a categorical diagnosis of clinically significant depression, our findings indicate that 34.7% of Goodville users in the USA suffer from depression. These results align with previous studies conducted during the COVID-19 pandemic, which also reported high rates of depression prevalence. In particular, it is noteworthy to compare our findings with those of Ettman et al. Two studies by Ettman et al. (2020, 2022) on depression prevalence in the US population during the pandemic have gained significant attention. Both studies utilized the PHQ-9 questionnaire to assess depressive symptoms. The first study (Ettman et al., 2020) evaluated symptoms at the beginning of the pandemic in March-April 2020 and reported a clinically significant depression rate of 27.8%. Another study (Ettman et al., 2022) estimated depression prevalence in spring 2021, where rates rose to 32.8% among US adults. Notably, our study found no statistically significant difference in the prevalence of clinically significant depression compared to these previous studies (34.7%).

Furthermore, it is of substantial interest to compare our study's findings with those of Tomitaka et al. (Tomitaka et al., 2018) regarding individual depression symptoms. Tomitaka et al. investigated the distribution of item responses and total scores using the PHQ-9 in the general US population. They utilized data from the 2013-2014 National Health and Nutrition Examination Survey. Figure 3 (Depression, n.d.) displays the distribution of depression symptoms based on Tomitaka et al.'s study.

Comparing the histograms in Figures 2 and 3, we observed a striking similarity in the ranking of depression symptoms based on prevalence frequency between our study and Tomitaka et al.'s study. Both studies identified low energy, sleep problems, and loss of interest as the most common symptoms. Consistent with both population studies, the rarest symptoms of depression in the US population, excluding suicidal thoughts (PHQ-9 item 9), were psychomotor disturbances, concentration difficulties, and low self-esteem. Notably, there was a substantial increase in the frequencies of more prevalent depression symptoms in the US population from 2013-2014 to 2021-2022.

Our comparative analysis leads us to conclude that the prevalence and severity of depressive symptoms significantly increased in the US population during the COVID-19 pandemic compared to the pre-pandemic period. However, the configuration of depressive symptoms remained largely unchanged, with no significant shift in the distribution of symptoms by frequency of occurrence. The increased severity of depression symptoms, coupled with the same frequency configuration, resulted in a notable rise in clinically significant depression, with somatic and motivational symptoms prevailing.

Furthermore, our study distinguishes itself from similar research in several aspects. The prevalence of depression was assessed by integrating the PHQ-8 questionnaire into a mobile game, enabling the collection of large-scale data. Over the period from April 2021 to May 2022, a total of 72,123 users from the US population completed the PHQ-8. In the initial data analysis, response patterns were examined for validity, and Rasch analysis was employed to validate the respondents' answers. Consequently, data from 54,856 US users were validated and utilized for assessing depression prevalence. The findings indicate a high rate of depression in the US population during the last two years of the COVID-19 pandemic (34.7%). Low energy, sleep problems, and loss of interest emerged as the most common depression symptoms in the US. The data obtained in this study exhibit complete consistency with the results of similar studies on depression prevalence in the US population, both in terms of overall prevalence of depression and the frequency distribution of specific symptoms. When compared to data from 2013-2014, the COVID-19 pandemic led to an increase in depression symptoms but did not alter their configuration. Somatic symptoms and loss of interest remained prevalent in the US population in 2021-2022, akin to the pre-pandemic period. However, during COVID-19, the severity of these symptoms, along with other depression-related symptoms, significantly escalated.

Our study demonstrates the high validity of integrating assessment tools into the Goodville app for gathering data on population mental health. Given the consistency between the mental health data from the Goodville app and data from well-established studies, information derived from the Goodville application can be employed to draw scientifically grounded conclusions about mental health and emotional well-being within extensive populations.

References

Astutik, E., Hidajah, A. C., Tama, T. D., Efendi, F., & Li, C. Y. (2021). Prevalence and determinants of depressive symptoms among adults in Indonesia: A cross-sectional population-based national survey. Nursing Forum, 56(1), 37–44. https://doi.org/10.1111/NUF.12508

Boone, W. J. (2016). Rasch analysis for instrument development. Life Sciences Educ., 15(4), 1–7.

Boone, W. J., Staver, J. R., & Yale, M. S. (2014). Rasch Analysis in the Human Sciences. In Book. https://doi.org/10.1007/978-94-007-6857-4

Depression. (n.d.). Retrieved July 25, 2022, from https://www.who.int/news-room/fact-sheets/detail/depression

Dhingra, S. S., Kroenke, K., Zack, M. M., Strine, T. W., & Balluz, L. S. (2011). PHQ-8 Days: A measurement option for DSM-5 Major Depressive Disorder (MDD) severity. Population Health Metrics, 9(1), 1–8. https://doi.org/10.1186/1478-7954-9-11/TABLES/6

Ettman, C. K., Abdalla, S. M., Cohen, G. H., Sampson, L., Vivier, P. M., & Galea, S. (2020). Prevalence of Depression Symptoms in US Adults Before and During the COVID-19 Pandemic. JAMA Network Open, 3(9), e2019686–e2019686. https://doi.org/10.1001/JAMANETWORKOPEN.2020.19686

Ettman, C. K., Cohen, G. H., Abdalla, S. M., Sampson, L., Trinquart, L., Castrucci, B. C., Bork, R. H., Clark, M. A., Wilson, I., Vivier, P. M., & Galea, S. (2022). Persistent depressive symptoms during COVID-19: a national, population-representative, longitudinal study of U.S. adults. The Lancet Regional Health - Americas, 5, 100091. https://doi.org/10.1016/j

Green, K., & Frantom, C. (2002). Survey development and validation with the Rasch model. International Conference on Questionnaire Development, Evaluation, and Testing, 1–30.

Grønli, O. K., Bramness, J. G., Wynn, R., & Høye, A. (2022). Depressive symptoms in the general population: The 7th Tromsø Study. Journal of Affective Disorders Reports, 8, 100322. https://doi.org/10.1016/J.JADR.2022.100322

Kessler, R. C., Andrews, G., Colpe, L. J., Hiripi, E., Mroczek, D. K., Normand, S. L. T., Walters, E. E., & Zaslavsky, A. M. (2002). Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med, 32(6), 959–976. https://doi.org/10.1017/s0033291702006074

Kroenke, K., Strine, T. W., Spitzer, R. L., Williams, J. B. W., Berry, J. T., & Mokdad, A. H. (2009). The PHQ-8 as a measure of current depression in the general population. Journal of Affective Disorders, 114(1–3), 163–173. https://doi.org/10.1016/j.jad.2008.06.026

Tomitaka, S., Kawasaki, Y., Ide, K., Akutagawa, M., Yamada, H., Ono, Y., & Furukawa, T. A. (2018). Distributional patterns of item responses and total scores on the PHQ-9 in the general population: Data from the National Health and Nutrition Examination Survey. BMC Psychiatry, 18(1), 1–9. https://doi.org/10.1186/S12888-018-1696-9/FIGURES/3

Wright, B. D., & Masters, G. N. (1990). Computation of OUTFIT and INFIT statistics. Rasch Measurement Transactions, 3(4), 84–85.

Wright, B., & Linacre, M. (1980). Reasonable mean-square fit values. 1–3.