The entrepreneur well-being check: a screener for entrepreneur mental health and well-being Open Access

Entrepreneurs' mental health and well-being (MWB) reflects a continuum ranging from mental illness to positive well-being (WHO, 2014), in the context of work engagement (Stephan, 2018). No MWB screening instruments are tailored for entrepreneurs. Deployment of an entrepreneur-focused screener could support timely referral to mental health care. The Entrepreneur Well-being Check (EWC) is a brief screening instrument developed to fill this gap.

A large literature supports the public health benefits of screening tools, including anxiety and depression measures to support employer-based mental health screening and intervention programs (Strudwick et al., 2023). While entrepreneurs lack comparable screening and intervention protocols, employer-based models could be adapted for entrepreneurs (Williamson et al., 2021). Employee screening measures, though, do not assess well-being and occupational functioning (Martin et al., 2020). Work is essential to the mental health and well-being of entrepreneurs because it is a channel through which they find meaning while expressing their identity (Stephan, 2018). Therefore, unlike screeners used for employees, the EWC also addresses well-being and occupational functioning.

Screening among entrepreneurs, and evidence-based interventions are warranted because mental health concerns (Hunt et al., 2022) and diminished well-being (Cocker et al., 2013) are common and related to poor occupational functioning (Stephan, 2018; Williamson et al., 2021). Large-scale representative studies and reviews consistently indicate that mental illness, psychological distress, and depression are common among self-employed samples (Willeke et al., 2021; Reid et al., 2018). The severity of these concerns is reflected in a finding that among 1,049 probability-sampled entrepreneurs, 3% reported having made suicide attempts, and 1.7% reported a history of psychiatric hospitalization (Hunt et al., 2022; Johnson et al., 2018), and in prior reports of elevated suicidality among self-employed persons (Min et al., 2019).

The high rates of mental health concerns reflect self-selection into entrepreneurship and the stresses of entrepreneurship. Regarding self-selection, for example, people with pre-existing conditions, such as attention-deficit/hyperactivity disorder (ADHD), are more likely to pursue entrepreneurship (Mannuzza et al., 1993). In a large U.S. representative sample, self-reported mental health diagnosis predicted the decision to become self-employed (Bogan et al., 2022).

Regarding adverse impacts, the occupational mental health hazards of entrepreneurship include precarity, highly-concentrated personal risk, low income, time pressure, long working hours, role ambiguity, work-family conflict, isolation, and high business failure rates (Cocker et al., 2013; Lechat and Torrès, 2017; Martin et al., 2020). Other pervasively entwined mental health issues include insufficient sleep (Williamson et al., 2019), burnout (Lechat and Torrès, 2016), and stress. A meta-analysis of 38 studies found that hindrance stressors (which constrain or thwart goal attainment) adversely affect psychological and physical well-being among entrepreneurs (Lerman et al., 2021).

Studies of large entrepreneur samples show that these hazards are linked to psychological distress, depression (Cubbon et al., 2021), and exit (Hessels et al., 2018). Business failure, bankruptcy, and involuntary exit from self-employment are common and sometimes associated with enduring serious mental health conditions (Nikolova et al., 2021).

Well-being and entrepreneurial performance are interconnected (see meta-analysis by Lerman et al., 2021) and bidirectional (Singh et al., 2007). The adjacency of well-being and performance reflects the complex interaction of mental health symptoms, their antecedents, and their occupational effects (Stephan et al., 2023). For example, success (antecedent) may cause anxious distress (symptoms) due to increased demands (occupational effect) that exceed coping capacity. Symptoms, in turn, may diminish occupational functioning. For example, among small-to-medium sized enterprise (SME) owners/managers who reported high distress, 82.5% endorsed impaired occupational functioning, 50% reported presenteeism/productivity loss, and 38.7% reported absenteeism in the past month (Cocker et al., 2013). Functional impairments due to mental health issues can cause serious personal and business problems that spread through connected innovation networks and weaken the social ties that help businesses succeed (Stam et al., 2014; Westlund and Adam, 2010). As an example of this feedback loop, production delays that result from the impairments of a depressed entrepreneur may impact employees, customers, and investors. Resulting business consequences may then exacerbate the entrepreneur's depression.

Conversely, positive entrepreneur MWB associates with better entrepreneurial and firm performance, persistence, engagement, opportunity recognition and creativity. Though the relationship may be nuanced and curvilinear, on average, “happy entrepreneurs are productive entrepreneurs” (Fodor and Pintea, 2017; Stephan, 2018).

Employers incur significant costs associated with employee mental health issues, including absenteeism, presenteeism, workers' compensation and healthcare claims, and short-term disability (Goetzel et al., 2018). Internationally, mental health conditions are the leading cause of employer disability days lost (Joyce et al., 2016).

Employer-sponsored occupational mental health programs demonstrate significant effectiveness and return on investment. A meta-analysis of 140 workplace intervention studies for anxiety and depression identified several interventions that improved symptoms and occupational outcomes (Joyce et al., 2016). In a review of workplace mental health interventions, several were robustly associated with improved productivity, return to work rates, employee retention, job effectiveness, and reduced absenteeism (Wagner et al., 2016). Research has converged on best practices for workplace mental health interventions (Wu et al., 2021). A meta-analysis of 8 such studies found that screening alone did not significantly improve mental health or economic outcomes, but screening coupled with intervention improved mental health outcomes (Strudwick et al., 2023). Parallel findings were observed in 9 workplace randomized clinical trials (RCTs) targeting depression (Tan et al., 2014). Insights gained from many validated employer-based MWB interventions, and person-based recovery interventions, could be adapted for entrepreneurs (Williamson et al., 2021).

Despite this evidence, there are major gaps. Most workplace mental health intervention studies have relied on only a single mental health concern, such as depression, perceived stress, anxiety (Bhui et al., 2012), or occasionally, depressive and anxiety symptoms, in both their assessments and their intervention targets (Martin et al., 2009b). This is problematic because, as described below, research suggests that a broader range of MWB concerns are common among entrepreneurs. In one exception, 344 employees completed a 48-item screener to assess stress, trauma, depression, anxiety, alcohol/substance use, sleep disturbance, financial stress, and social support (Fragala et al., 2021). One-third of employees identified as high-risk took advantage of help services offered, highlighting that broad screening may identify substantial numbers who will seek care. Despite the strengths, we will describe multiple core entrepreneur concerns that were not covered by the scales used.

Entrepreneurs benefit from interventions that emphasize recovery, including sleep, mindfulness, exercise, coaching, and detaching from work (Williamson et al., 2021). Nonetheless, the evidence for MWB interventions among entrepreneurs is insufficient, which may contribute to the unavailability of evidence-based MWB support for SME owners (Martin et al., 2009a; Willeke et al., 2021, 2024). The few entrepreneur occupational mental health interventions that have been tested employ screeners and show promising results. For example, among entrepreneurs, cognitive behavioral therapy workshops reduced anxiety and depression symptoms and improved well-being (Saraf et al., 2019). In an RCT of SME owner/managers who were identified using anxiety and depression screeners, telephone-facilitated interventions, and to a lesser extent, psychoeducational mental health resources, related to significant decreases in distress (Martin et al., 2020).

As with employer-sponsored workplace interventions, these entrepreneur interventions have relied on general population screening measures, such as the Patient Health Questionnaire-Anxiety and Depression Scale (PHQ-ADS) and the WHO-5 (Saraf et al., 2019). Large studies have supported the sensitivity of the K6 (Bogan et al., 2022; Reid et al., 2018) and K10 (Cocker et al., 2013) scales among entrepreneurs, in that many entrepreneurs endorsed high levels of distress. These scales, though, were neither tailored for, nor validated with entrepreneurs. Although entrepreneur burnout scales have been used to screen business owners (Torrès et al., 2022), our goal is to create an entrepreneur-centric instrument to support much-needed RCTs of entrepreneur MWB interventions (Williamson et al., 2021) by capturing a broader range of entrepreneur concerns.

Despite the importance of screening for entrepreneur MWB, and for guidance on MWB intervention implementation (Williamson et al., 2021), no instruments cover the range of MWB concerns that are well-documented among entrepreneurs. This research sought to develop and validate a screener to effectively cover a broad spectrum of the mental health, well-being, and functional concerns commonly reported by entrepreneurs. Here, we present the process of developing the scale and its preliminary psychometric evidence.

This study was approved by Sterling Institutional Review Board (IRB).

Scale development was guided by standard steps of (1) construct definition, (2) operationalization of constructs, and (3) empirical tests of the validity of the item set (Lambert and Newman, 2023). For construct definition, we conducted an extensive literature review to identify key mental health, well-being, and occupational functioning concerns among entrepreneurs, drafted items to capture each core concern, and then tested the scale among entrepreneurs. We followed standard guidance for psychological measure development (Clark and Watson, 2019) to examine the psychometric qualities of the scale within entrepreneurs. Items were rated by entrepreneurs for content validity, clarity, and breadth. Then, in a second sample, we examined convergent and divergent validity of each item using a multi-trait matrix and tested the factor structure, internal consistency and test-retest reliability. We tested criterion validity by assessing whether EWC total scores significantly correlated with well-validated (longer) scales of mental health, occupational functioning, or well-being. Moreover, we tested broader construct (nomological net) validity by considering whether the EWC was significantly correlated with better business outcomes, based on established evidence linking mental health with entrepreneurial performance.

Scale development was supported by “engaged scholarship,” which reflects the reciprocal relationship between theory and real-world practice (Van de Ven and Johnson, 2006). During the scale development process, two authors delivered over 1,300 combined hours of direct therapeutic services to entrepreneurs. This enabled real-time theory testing and refinement based on “bench and bedside” inputs.

We systematically reviewed literature on three commonly-referenced, interrelated entrepreneur MWB domains: mental health, well-being, and occupational functioning. We sought to identify key aspects of entrepreneur MWB, and to confirm that there were no instruments to measure these facets of MWB among entrepreneurs.

Over 1,000 meta-analyses, reviews, and original studies were identified through NCBI, PubMed, PsycInfo, and Google Scholar databases. Keywords included mental health, well-being, and occupational functioning/performance/success term synonyms, along with self-employed* or entrepreneur*. Additional relevant studies were identified by reviewing the citations in meta-analyses and reviews. Only peer-reviewed articles were considered. Our goal was to capture the entire well-being continuum and to identify domains that were either commonly recognized as concerns among entrepreneurs, or that were validated as relevant to entrepreneurial outcomes.

Our review identified 9 common concerns among entrepreneurs that have been empirically validated as important to entrepreneurial function: well-being (thriving, life satisfaction, social functioning), occupational functioning (work role fulfillment, burnout, intent to quit), and mental health (anger/hostility, negative emotionality, substance use). After item review consultation with entrepreneur mental health experts, we added a tenth concern that was not originally identified; sleep impairment. Taken together, and guided by Hinkin (1998), we selected 10 issues to cover, assuming that at least six would be retained after factor analysis. A literature summary for each dimension is provided in the supplementary material.

For each item, we conducted literature review to identify validated public-domain single-item measures that could serve as potential EWC items. This yielded the widely-used and well-validated Single Item Life Satisfaction Scale (SILSS; Cheung and Lucas, 2014). We used deductive methods, relying on definitions of the concerns that had emerged in the literature review (Boateng et al., 2018; Lambert and Newman, 2023), to write 9 novel items to measure the focal concerns other than life satisfaction. We used the global domain sampling approach to develop single items reflecting each concern (Lambert and Newman, 2023).

A team including three scientists with extensive clinical experience collaboratively drafted and revised the 9 novel items to capture each of the related initial concerns. A 5-point response scale with verbal anchors was used to increase variability and reliability (Lambert and Newman, 2023). To further specify construct definition, we limited the reference period for each domain and concern to the last 30 days, to focus on acute emotional distress. This reflects the best practices in screening instruments for minimizing long-term recall, and supports future head-to-head comparisons between these instruments.

As recommended across scale development guidelines (Boateng et al., 2018; Lambert and Newman, 2023), we considered the content validity of items, drawing on the expertise of entrepreneurs. Twenty-two entrepreneurs from the US (M_age = 48.4; 72.7% male) were recruited through personal outreach, asked to evaluate the ten preliminary EWC items on clarity and relevance for entrepreneurs, and asked one open-ended question about what else would be most relevant to assess in the EWC. Responses indicated that the EWC is adequately broad, clear, and relevant (Table 1), hence, we retained all ten items. Overall scale relevance was rated as moderately high (M = 3.2, SD = 0.7). Based on qualitative participant feedback, we rephrased six items (thriving, social functioning, occupational functioning, burnout, intent to quit, and anger/hostility). See Table S1 for feedback and item edits.

We recruited an international sample of 676 participants between May and October 2024 through venture capital firms, university-based entrepreneurship programs, entrepreneur conferences and media, incubators, and organizations engaged with growth-oriented entrepreneurs, who sent emails to potential study participants and posted social media and newsletter announcements. Inclusion criteria were assessed with pre-screening questions: being a company founder or co-founder, > 18 years old, fluent in English, and a resident of the United States, United Kingdom, European Union, Canada, or Israel.

Informed consent, demographic and business performance information, the EWC, baseline psychometric measures, and four embedded attention-check items were administered online using RedCap (Harris et al., 2019). Follow-up assessment was conducted 6 months later. Estimated survey completion time was 45 min for baseline and 15 min for follow-up. Participants received 25 USD for survey completion at baseline and follow-up.

Scales are scored as the sum of items unless otherwise stated (see Table 2 for score ranges). Internal consistency of each scale was adequate (Table 2).

EWC: The preliminary EWC includes ten items. Higher scores indicate better well-being and mental health in the last month. Items vary in being positive or negative, thus, the response format varied across items (see Table A1 in the  Appendix and supplementary material). Items are averaged.

Measures to Assess the Validity of Specific EWC Items. We selected one or two public domain self-report instruments with well-replicated, strong psychometric properties to validate each EWC item. As higher EWC scores indicate more well-being, we expected EWC items to correlate positively with the well-being scales, but negatively with scales where a higher score was indicative of problems.

WHO-5 Well-being Index (WHO-5: Topp et al., 2015): five-item scale to assess eudaimonic and hedonic well-being.

Flourishing Scale (FS: Diener et al., 2010): eight-item scale to assess well-being, and psychological flourishing.

Satisfaction with Life Scale (SWLS: Diener et al., 1985): five-item scale to assess life satisfaction.

PROMIS Global Mental and Physical Health Scale (Hays et al., 2009): single-item Global05 (“In general, how would you rate your satisfaction with your social activities and relationships?”) and Global09 (“In general, please rate how well you carry out your usual social activities and roles; this includes activities at home, at work, and in your community, and responsibilities as a parent, child, spouse, employee, friend, etc.”) to assess social functioning.

Entrepreneurial Self-Efficacy Scale (ESE: Zhao et al., 2005): four-item scale to assess self-efficacy in entrepreneurial endeavors. Items are averaged.

Burnout Measure – Short Form (BMS-10: Malach-Pines, 2005): 10-item scale to assess burnout. Items are averaged.

Turnover Intention Scale (TIS: Michaels and Spector, 1982): three-item scale to assess the intention to quit current entrepreneurial endeavors. Items are averaged.

Kessler Psychological Distress Scale 6 (K6: Kessler et al., 2003): six-item scale to assess past-month negative emotionality (i.e. anxious and depressive emotions).

Buss-Perry Aggression Questionnaire – Short Form (BPAQ-SF: Diamond and Magaletta, 2006): 12-item scale to assess anger, aggression, and hostility. Items are averaged.

PROMIS Sleep-Related Impairment – Short Form (Yu et al., 2011): eight-item scale to assess past-week sleep-related difficulties. One positively worded item is reverse-coded.

CAGE Questionnaire Adapted to Include Drug Use (CAGE-AID: Williams, 2014): four-item scale to assess substance use problems.

Measures to Test the Criterion Validity of the EWC Total Score. We selected the following measures to assess EWC total score criterion validity (Lambert and Newman, 2023). Because higher scores on the mental health scales indicate more distress, we expected negative correlations with the EWC total score. Furthermore, positive business outcomes were expected to correlate positively with the EWC total score.

DSM-5 Self-Rated Level-1 Cross-Cutting Symptom Measure – Adults (DSM-5 Cross-cutting symptoms: Narrow et al., 2013): 23-item scale designed to assess the likelihood of one or more mental health conditions. Items cover past two-week symptoms (i.e. depression, anger, mania, anxiety, somatic symptoms, suicidal ideation, psychosis, sleep problems, memory, repetitive thoughts and behaviors, dissociation, personality functioning, substance use). For domains comprising multiple items, the highest score is considered.

Mood Disorder Questionnaire (MDQ: Hirschfeld et al., 2000): to assess mania/hypomania. We focused on the 13-item symptom severity score.

Adult ADHD Self-Report Scale (ASRS: Kessler et al., 2005): 18-item scale to assess ADHD. The response scale ranged from 0 to 4 and is recoded to provide 1 point for each item rated > 2 or 3, depending on the item.

Primary Care PTSD Screen for DSM-5 (PC-PTSD-5: Prins et al., 2015): six-item scale to assess post-traumatic stress disorder (PTSD). Participants are asked if they have had trauma; those who endorse this are asked to rate five PTSD symptoms items.

Business Outcome Items: Financing raised, number of full-time employees, and top-line annual revenue were assessed as raw numbers. Participants were asked if they had protected intellectual property (IP; copyright trademark, granted or provisional patent); adverse business outcomes (insolvency of magnitude that the firm could no longer attract debt or equity funding, choosing to stop because the business was not economically viable, involuntary exit from self-employment into unemployment or a job, or bankruptcy); money invested in one's company; development of new products, services or processes in the last year, assessed as dichotomous variables. Number of businesses operating on a scale ranging from “none” to “four or more”. Business survival was calculated as the number of years of operation. Past-year profit/loss margin; and growth/contraction in revenue, customers, and employees were assessed with a slider ranging from “+200% or more” to “–200% or more”.

We excluded 171 records that appeared to be completed by bots (e.g. suspiciously rapid timing, sequential emails, completion of hidden Redcap items). We excluded 173 participants who dropped out before completing the EWC, 8 participants with probable careless responding (i.e. two or more failed attention check questions, e.g. “for the quality of our data, please select strongly agree”), and 10 participants who completed more than 18% of survey pages too quickly (i.e. less than 2 s per item; Bowling et al., 2023). Excluded participants did not significantly differ from the analysis sample demographically (all p's > 0.05). Manually screening for geometric patterns (e.g. 14 or more identical responses in a row) and outlier scores (Z >|3|) led to exclusion of two instruments (FS and SWLS) in one participant. After exclusions, the analysis sample included 314 participants (age range: 21–80 years) from multiple countries with companies in 8 economic sectors (see Table 2 and 3 for additional demographic information).

181 participants completed the six-month follow-up. We excluded five participants who completed two out of four survey pages too quickly, four participants who dropped out before completing the EWC, and one participant with probable careless responding. Accordingly, the follow-up analysis sample size was 171.

We imputed missing item-level data to form scale totals when at least 60% of scale items were complete. Variable distributions were examined for adequate variability and normalcy (Clark and Watson, 2019). We selected transformations for extremely skewed or kurtotic variables that most successfully normalized the distribution.

As recommended by Lambert and Newman (2023), we examined convergent and divergent validity of specific items through a multi-trait matrix approach, which comprised a Pearson correlation matrix of each EWC item with measures selected to cover the same constructs, as compared to measures of other constructs (see Table 4 to view expected links in italic). We applied Bonferroni adjustments for the number of convergent validity correlations tested (p = 0.05/12 = 0.0042).

Subsequently, we performed confirmatory principal component analyses (PCA) with oblimin rotation to test the predicted structure of the EWC outlined above (Lambert and Newman, 2023). We evaluated the model goodness-of-fit by using conventional cut-off values of model fit indices: χ2 (p > 0.05), comparative fit index (CFI) > 0.90, and root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) < 0.08 (Byrne, 1994). Because fit was inadequate, we conducted an exploratory PCA. We also applied exploratory PCA to identify the number of separable dimensions in business outcome indices and across the well-being and mental health validation measures. We assessed Kaiser-Meyer-Oklin (KMO) to examine sampling adequacy (KMO >0.5; Kaiser, 1974) and Horn's Parallel Analysis to determine the number of factors to be extracted (Horn, 1965). For all PCAs, we iteratively excluded items with low communality (i.e. <0.40) starting with the item with the lowest communality. After identifying the number of factors to retain, we iteratively excluded items with high cross-loading to achieve a simple structure (i.e. items assess only one construct; Lambert and Newman, 2023), starting from those with the strongest cross-loading.

For the final EWC scale, internal consistency was assessed as McDonald's ω (Malkewitz et al., 2023). Test-retest reliability was assessed as the Pearson correlation between the baseline and 6-month follow-up EWC scores (Boateng et al., 2018).

To assess criterion validity, we analyzed Pearson correlations of the EWC total with the DSM-5 Cross-cutting symptoms, mania/hypomania, ADHD, and PTSD, and for nomological validity, we considered correlations with business outcome indices. We applied Bonferroni adjustments for the number of correlations tested (0.05/18 = 0.0028 as the significance threshold).

To test potential confounds, we considered whether demographic variables (i.e. age, gender, education, and number of children) were correlated with the EWC, psychopathology, and business outcomes. Where correlations were significant with both the EWC and the other measure, we provide partial correlation analyses controlling for those demographic or economic variables. Supplemental analyses included regression models of whether the correlations of EWC with validity indicators varied by potential confounds (Table S2), the validity of the sleep item (Table S3), the replication of EWC total score criterion validity at the 6-month follow-up (Table S4), and the 6-month test-retest of the DSM-5 Cross-cutting symptoms (Table S5).

Across all variables, 0.3–10.8% of data were missing. Little's Missing Completely at Random test (Little, 1988) indicated that the data were missing completely at random, χ2(30,675) = 30,507.36, p = 0.750. Descriptive indices are presented in Table 2.

Our analyses were pre-registered on Open Science Framework (https://osf.io/ym2sa/?view_only=c67541a7565e4adf860cadade1308d2e). We deviated from planned analyses in three ways. We eliminated analyses of the three expected EWC subscales (mental health, occupational functioning and well-being) because factor analysis only supported the formation of a unidimensional total EWC score. We used continuous scores from the DSM-5 Cross-cutting symptoms instead of binary scores to increase statistical power (Cohen, 2013). We excluded responses of <2 seconds-per-item per page, based on recent empirical analyses (Bowling et al., 2023).

Financing raised was transformed into five categories (i.e. 0, 1–100,000, 100,001–1000,000, 1,000,001–5000,000, >500,000). The FS score was flipped by subtracting the total score from 57 and then log transformed. Number of employees, number of businesses operating, number of children, DSM-5 Cross-cutting memory and dissociation symptoms were log transformed. Revenue was winzorized by replacing extreme outliers with 1,000,000. Business survival was square-root transformed. DSM-5 Cross-cutting symptoms of suicidality and psychosis were excluded from further analyses due to low variability and extreme kurtosis (34.38 and 29.16, respectively).

In the exploratory PCA of the validation measures, the ESE scale was excluded due to low communality. PCA identified two factors surpassing Horn's Parallel Analysis threshold, which accounted for 55.2% of the variance. However, cross-loading of multiple items was high. Accordingly, we retained the separate validation scale totals, so as to provide more detailed information.

We conducted exploratory PCA with 12 business outcome items. Iterative elimination of items with high cross-loading (IP awards received and financing raised) or low communality (new product development and adverse business outcome) resulted in the retention of eight business outcome items. Eigenvalue of Factors 1 and 2 (2.57 and 1.90, respectively) surpassed Horn's Parallel Analysis thresholds (Factor 1 = 1.25; Factor 2 = 1.16). We excluded Factor 3 because the eigenvalue (1.01) did not exceed Horn's Parallel Analysis threshold of 1.09. This resulted in two statistically separable factors of “business growth” and” business size and stability”, r = 0.023, that accounted for 55.8% of the variance. Factor loadings are shown in Table S6. KMO suggested adequacy of sampling, 0.65. Business outcome items were z-transformed before computing the factor mean scores.

All correlations were in the expected direction (Table 4). Convergent validity correlations were at least moderate in size and significant at the Bonferroni-adjusted significance threshold (p < 0.0042), indicating good convergent validity. Correlations for the life satisfaction, social functioning, and intent to quit items with the longer validity scales were high (>0.60). Contrary to expectations, discriminant validity correlations were not substantially lower. That is, items designed to assess mental health symptoms showed moderate correlations not only with the symptom validation scales but also with occupational functioning and well-being scales. Mental health, occupational functioning, and well-being items were moderately interrelated. Nonetheless, correlations were highest between EWC items and validation measures that assessed the same construct, except for the WHO-5 and FS scales, which were more strongly correlated with EWC life satisfaction and social functioning instead of thriving.

Tests of the number of dimensions are widely recommended in scale development (Boateng et al., 2018; Lambert and Newman, 2023). We first tested confirmatory PCA with three factors of well-being, occupational functioning, and mental health. We removed three items with low communality (substance use, anger/hostility, and intent to quit). Contrary to the preregistered expectations, it showed poor model fit, χ2(32) = 70.24, p < 0.001; CFI = 0.77, despite adequate RMSEA = 0.06 and SRMR = 0.07. In the subsequent exploratory PCA we excluded the intent to quit item due to low communality, and we identified two factors that surpassed Horn's Parallel analysis thresholds. However, three items exhibited high cross-loading. Thus, we constrained the PCA to one factor, excluding substance use and anger/hostility due to low communality. All seven items loaded on a single factor with eigenvalue of 3.48, surpassing Horn's Parallel Analysis threshold of 1.21. Final items are shown in Table A1 in the  Appendix. Factor loadings are shown in Table S7. KMO suggested adequacy of sampling, 0.84. The single factor accounted for 49.7% of the variance. The total EWC total score was computed as the mean of the 7 EWC items. Descriptive indices are shown in Table 2. Internal consistency of the EWC total score was McDonald's ω = 0.82.

Pearson correlations of the EWC score with psychopathology, business outcome indices, adverse business outcomes, and IP awards received are presented in Table 5. The Bonferroni adjusted significance threshold (p < 0.0028) was surpassed for EWC correlations with DSM-5 Cross-cutting symptoms for depression, anxiety, sleep problems, personality functioning, anger, repetitive thoughts and behaviors, somatic symptoms, memory, and dissociations, with separate measures of PTSD and ADHD and with the business growth index. Null results were observed for mania and substance use. We observed parallel correlations in our 6-month follow-up assessment (see Table S4).

The EWC score was significantly correlated only with age (r = 0.13, p = 0.023), other r's < |0.08|, p > 0.141. Age was also correlated with multiple mental health and business outcome indices (Table 5). Accordingly, for variables that significantly related to age, we provide the partial correlation of EWC with the mental health and business outcome indices, controlling for age (Table 5). The bivariate significant correlations remained significant when controlling for age, except for dissociation (which was rare in this sample, and may have limited statistical power) and business growth.

The regression models shown in Table S2 suggested that none of the potential confound variables interacted with the EWC in predicting mental health or business outcomes.

The EWC total score was stable over time (t = −0.83, p = 0.407) and 6-month test-retest reliability was r = .66, p < 0.001. These indices were as strong or stronger than those for the DSM5 Cross-cutting symptoms (see Table S5).

Screeners used in employee occupational mental health programs often focus on single mental health dimensions and were not developed to cover entrepreneur MWB concerns. By contrast, the Entrepreneur Well-being Check (EWC) is explicitly designed to assess a broad range of mental health, well-being and occupational functioning issues shown to be particularly relevant for entrepreneurs. Here, we validated the EWC in a diverse international entrepreneur sample, using pre-registered hypotheses and analyses.

Consistent with our aim of creating a broader screener, the EWC scale showed criterion validity not only in relation to depression and anxiety, but also for ADHD, PTSD and sleep problems, and modest correlations with many other symptoms. Most effects were sustained when considering potential covariates. Thus, a brief instrument can be used to detect a broad range of common mental health-related concerns among entrepreneurs. The only mental health domains not related significantly to the EWC score were mania (DSM-5-Cross-cutting symptoms and MDQ score) and substance use— less common problems in our sample – which are omitted from the final EWC items.

The entrepreneur MWB construct includes mental health, well-being, and occupational functioning. Rather than reflecting three separable dimensions as expected, EWC items were captured by a single dimension. The unidimensional nature of this construct supports recent entrepreneur MWB theory emphasizing the close association of mental health symptoms with well-being and occupational functioning among entrepreneurs (Stephan, 2018).

Altogether, our 7-item scale (Table A1 in the  Appendix) provides a valid index of mental health concerns among entrepreneurs. Our scale covers more breadth, and does so more concisely, than previously available scales or combinations of scales.

The EWC test-retest reliability was modest but still higher than the DSM5 Cross-cutting measure. Nonetheless, given fluctuations in entrepreneurial demands and related MWB, we recommend repeated assessments for those who aim to detect and intervene around entrepreneur MWB.

The EWC is affordable and accessible. It can be administered easily across settings by non-clinical practitioners, coaches, and therapists, and digitally to provide a simple gateway to care and to assess MWB issues that could influence performance. Positive results could be paired with immediate offers of self-care and referral resources.

At the social and system level, mental health screening can help governmental small business agencies provide better guidance to business owners. The EWC can be deployed by large ecosystem stakeholders that endeavor to improve the well-being and mental health of affiliated entrepreneurs, such as development banks, venture capital associations, and entrepreneurship networks. Population-level EWC scores can be used by researchers and agencies to assess the MWB impact of economic shocks, and to inform occupational well-being and mental health policy and programs designed for entrepreneurs. For example, population EWC score distributions can be analyzed to identify distressed entrepreneur populations (e.g. refugee entrepreneurs) that may benefit from targeted interventions. Future studies can determine clinical cut-points and ranges for flagging mild, moderate and severe risk of impairment.

Broader commercial benefits may accrue to innovation ecosystems that incorporate the EWC into coordinated programs of screening, detection, and early intervention. The performance and spillover benefits of entrepreneur well-being may accrue not only to the entrepreneur, but also to related ecosystem stakeholders including their employees, suppliers, channel partners, customers and investors.

Due to its brevity, the EWC can be used by incubators, accelerators and entrepreneurship educators to normalize mental health differences, to help nascent entrepreneurs develop self-awareness about their own well-being, learn about the emotional impact of entrepreneurship, and sustain well-being while building businesses.

There is currently insufficient evidence upon which to base occupational MWB interventions for entrepreneurs (Martin et al., 2020; Williamson et al., 2021). The EWC could help to address this deficiency by facilitating research on entrepreneur-centric interventions.

Beyond screening and detection, there are gaps in the other core pillars of population health management programs for entrepreneurs. These include primary prevention, secondary early intervention, tertiary treatment and mental health maintenance services (Purtle et al., 2020; Steenkamer et al., 2017), even though highly effective medication and psychological therapies are widely available. Entrepreneurs, and perhaps their companies and ecosystems, are likely to benefit from the full range of occupational health management supports available to most other employed people in industrialized countries (Khan et al., 2022).

We acknowledge multiple limitations. Entrepreneurs with mental health concerns may have been more likely to participate. Our sample showed low levels of aggression, suicidality, and substance use, which may have artificially dampened the correlations of those items with the EWC total score. Less severe anger item content could improve sensitivity. Our item regarding intent to quit did not perform well, perhaps because we merged problematic (e.g. “quiet quitting”) and successful scenarios (e.g. quitting due to a good exit). Although we used careful data validation to combat reporting issues, we relied entirely on self-report measures; common methods bias may artificially inflate the effect sizes observed. We conducted many analyses; concerns about Type I errors are somewhat relieved by the match of findings to pre-registered hypotheses and our use of strict Bonferroni significance thresholds. A key goal, though, is to replicate the validity of the EWC using a multi-method approach.

The EWC has sound psychometric properties. It detects a broad range of entrepreneur MWB concerns. Due to its brevity, the EWC may be used to screen entrepreneurs and identify those who could benefit from referral for mental health care. The EWC could be used to develop evidence that supports entrepreneur-focused occupational well-being and mental health interventions. The EWC can thus contribute to reducing morbidity and mortality among entrepreneurs while improving their well-being and occupational functioning.

Sterling IRB reviewed our study protocol and determined that it was exempt from IRB review (IRB ID: 11,905-SJohnson01).

The authors gratefully acknowledge unrestricted research funding by Flourish Ventures, the Business Development Bank of Canada, Intel Ignite (now Ignite), and the McCall Family Foundation. Funders had no role beyond financial support. We also thank the ecosystem stakeholders that supported recruitment, the participating entrepreneurs, and our colleagues who reviewed this manuscript.

The supplementary material for this article can be found online.