Notes on Anxiety of Health Care Workers in Radiology During COVID-19

Background

• COVID-19 outbreak increased infection risk for health care workers, especially in radiology where frontline staff interact with suspected patients and face sudden workload surges.
• Anxiety can harm physical/mental health, work efficiency, and patient safety; necessary to assess anxiety levels and identify risk/protective factors in high-exposure radiology workers.

Aims and scope

• First psychological impact survey among radiology department workers with high exposure risk during the early COVID-19 outbreak in China.
• Objectives: determine prevalence of psychological anxiety; identify risk factors and protective factors; examine potential role of resilience.
• Hypothesis/rationale: resilience may mitigate anxiety; timely assessment can guide interventions.

Definitions and measures used

• SARS-CoV-2 and COVID-19: novel coronavirus outbreak beginning December 2019; global spread with public health emergency implications.
• Anxiety scales used:
  • Self-Rating Anxiety Scale (SAS): 20 items, 7-day period; rough score R; standard score S = ⌊1.25 × R⌋; range $S \,\in\, [25, 100]$; higher scores indicate higher anxiety. Categories: no anxiety (<50), mild (50–59), moderate (60–69), severe (≥70). Cronbach’s alpha in this study: $\alpha = 0.839$.
  • Chinese Connor-Davidson Resilience Scale (CD-RISC): 25 items, 0–100 scale; higher scores indicate higher resilience. Cronbach’s alpha in this study: $\alpha = 0.961$.
• Other concepts discussed: resilience as a protective factor; negative emotions transfer and the impact of others’ emotions on individuals; importance of PPE in reducing anxiety.

Study design and participants

• Design: multicenter cross-sectional survey, anonymous online questionnaire via WeChat.
• Ethical approval: West China Hospital, Sichuan University. Informed consent obtained.
• Timeframe: February 7–9, 2020.
• Sampling frame: radiology departments in 32 public hospitals in Sichuan Province, China.
• Inclusion criteria: age ≥18; nurses and technicians in radiology; consent.
• Exclusion criteria: substance abuse/dependence; history of mental illness per CCMD-3/DSM-IV; current brain lesion or serious physical disease; invalid questionnaire (e.g., completion time <2 min or extreme answer patterns).
• Sample size planning: Kendall’s formula for multiple regression variables, augmented by 10% buffer; target minimum ~380.
• Final: 377 recruited; 364 valid questionnaires (96.6% of 377; 364/377).

Demographic and occupational characteristics (N = 364)

• Gender: Male 41.2% (n=150); Female 58.8% (n=214)
• Age distribution: <30 years 37.4% (n=136); ≥30 years 62.6% (n=228)
• Work experience: <10 years 48.4% (n=176); ≥10 years 51.6% (n=188)
• Education level: College or below 31.6% (n=115); Bachelor 66.5% (n=242); Postgraduate or above 1.9% (n=7)
• Marital status: Unmarried 25.5% (n=93); Married 72.8% (n=265); Divorced 1.7% (n=6)
• Job function: Nurse 32.7% (n=119); Technician 67.3% (n=245)
• Hospital classification: Grade 3A 65.9% (n=240); Grade 3B 30.8% (n=112); Grade 2A or below 3.3% (n=12)
• Residence: Live alone 13.2% (n=48); Live with roommate(s) 8.8% (n=32); Live with family 78.0% (n=284)
• Presence of suspected symptoms in participant: Yes 7.1% (n=26); No 92.9% (n=338)
• Presence of suspected symptoms in family members: Yes 5.5% (n=20); No 94.5% (n=344)
• Contact with confirmed/suspected patients at work: Yes 48.6% (n=177); No 51.4% (n=187)
• Availability of adequate protective materials: Extreme shortage 13.2% (n=48); Mild shortage 37.6% (n=137); Sufficient 28.0% (n=102); Abundant 21.2% (n=77)
• Knowledge about COVID-19: Insufficient 36.0% (n=131); Sufficient 64.0% (n=233)
• Susceptible to emotions and behaviours of people around them: Yes 20.9% (n=76); No 79.1% (n=288)
• Fear of inability to pay rent or mortgage: Yes 12.4% (n=45); No 87.6% (n=319)
• Fear of an uncontrollable epidemic: Yes 86.8% (n=316); No 13.2% (n=48)
• Psychological resilience (CD-RISC): Low <50 16.2% (n=59); High ≥50 83.8% (n=305)

Anxiety outcomes

• Overall anxiety score (SAS) mean and dispersion: $\text{Mean} = 44.28, \text{SD} = 8.93$
• Anxiety categories among participants: No anxiety $n=279\ (76.6\%)$; Mild anxiety $n=63\ (17.3\%)$; Moderate anxiety $n=19\ (5.2\%)$; Severe anxiety $n=3\ (0.8\%)$
• Distribution table summary (group means):
  • No anxiety group: mean $\approx 40.29\ (SD\ \approx 4.70)$
  • Mild anxiety group: mean $\approx 54.31\ (SD\ \approx 2.64)$
  • Moderate anxiety group: mean $\approx 63.55\ (SD\ \approx 3.01)$
  • Severe anxiety group: mean $\approx 82.50\ (SD\ \approx 2.17)$

Univariate analyses (factors associated with anxiety)

• Compared groups by various factors (t-tests or ANOVA as appropriate) identified significant associations:
  • Age: older age associated with higher anxiety (P = 0.011)
  • Gender: differences observed (P = 0.004)
  • Job function: nurses had higher anxiety than technicians (P = 0.001)
  • Availability of protective materials: significant differences across levels (P = 0.001)
  • Presence of suspected symptoms in participant: higher anxiety when symptoms present (P < 0.001)
  • Presence of suspected symptoms in family: not significant (P = 0.411)
  • Contact with confirmed/suspected patients at work: not significant (P = 0.308)
  • Knowledge about COVID-19: insufficient knowledge linked to higher anxiety (P = 0.049)
  • Susceptibility to emotions/behaviors of people around them: significant (P = 0.002)
  • Fear of uncontrollable epidemic: significant (P = 0.018)
  • Psychological resilience (CD-RISC): low resilience linked to higher anxiety (P < 0.001)
• Note on direction: higher anxiety observed with presence of participant symptoms, susceptibility to others’ negative emotions, older age, nursing role, PPE shortages, and fear of uncontrolled outbreak; resilience showed the opposite pattern (higher resilience associated with lower anxiety).

Multivariate linear regression analysis (predictors of anxiety)

• Model: anxiety score as the dependent variable; predictors entered from significant univariate factors.
• Overall model statistics:
  • $R^2 = 0.251, \ \text{Adjusted } R^2 = 0.238$
  • F = 19.945,\ P < 0.001
• Significant predictors and their effects (B, SE, standardized coefficient b, t, P):
  • Presence of suspected symptoms in participant: B = -7.188,\ SE = 1.620,\ b = -0.208,\ t = -4.436,\ P < 0.001
  • Susceptibility to emotions and behaviours of people around them: $B = 2.804,\ SE = 1.027,\ b = 0.128,\ t = 2.731,\ P = 0.007$
  • Job function (nurse vs. technician): $B = -2.085,\ SE = 0.902,\ b = -0.110,\ t = -2.311,\ P = 0.021$
  • Psychological resilience (low vs. high): B = -8.454,\ SE = 1.114,\ b = -0.349,\ t = -7.590,\ P < 0.001
  • Availability of adequate protective materials: $B = -1.125,\ SE = 0.437,\ b = -0.122,\ t = -2.572,\ P = 0.011$
  • Age (≥30 years vs <30): $B = 1.929,\ SE = 0.884,\ b = 0.105,\ t = 2.181,\ P = 0.030$
• Interpretation of coefficients (direction and meaning):
  • Presence of suspected symptoms in participant tends to lower the predicted anxiety score in the model by 7.188 points when coded as the reference category (note: actual group means show higher anxiety with symptoms; the coding in the model reflects a particular reference scheme). In practice, having symptoms is associated with higher anxiety.
  • Greater susceptibility to negative emotions among people around them increases anxiety by about 2.80 points per unit of the susceptibility measure.
  • Being a nurse (vs technician) is associated with a decrease of about 2.09 points in anxiety in this model; however, nurses showed higher anxiety in univariate results, suggesting interaction effects or coding differences.
  • Higher resilience (CD-RISC) strongly lowers anxiety; low resilience adds about 8.45 points to predicted anxiety.
  • Shortages/adequacy of PPE: more shortages correlate with higher anxiety; each category change in PPE adequacy reduces predicted anxiety by about 1.13 points when adequate protections are present.
  • Age: older workers (≥30) tend to have higher anxiety by about 1.93 points than younger workers, controlling for other factors.
• Note: The overall R^2 indicates that about 25.1% of the variance in anxiety scores was explained by the model; remaining variance may be due to unmeasured factors (environment, family support, workload, etc.).

Discussion and interpretation of findings

• Overall level of anxiety was high at the beginning of the outbreak among radiology workers with high exposure risk, but most remained within normal limits (i.e., not in severe anxiety range).
• Age-related finding: older workers (>30 years) showed higher anxiety, consistent with Ebola-era findings where older medical workers faced greater anxiety possibly due to heavier family responsibilities and risk judgments.
• Job role: nurses exhibited higher anxiety than technicians in univariate analyses, likely due to more direct patient contact and broader responsibilities (drug administration, patient management, environmental disinfection).
• Symptoms in participants: presence of suspected COVID-19 symptoms elevated anxiety and suggests the value of prompt SARS-CoV-2 testing and symptom screening to reduce psychological burden.
• PPE shortages: insufficient protective materials were associated with higher anxiety, underscoring the importance of reliable PPE supply to maintain morale and performance.
• Susceptibility to negative emotions around others: those who are more empathetic or sensitive to others’ emotions experienced greater anxiety, possibly via emotional contagion and negative emotion transfer.
• Resilience as a protective factor: higher CD-RISC scores correlated with substantially lower anxiety; resilience helps maintain adaptability and positive expectations during emergencies.
• Practical implications: Routine mental health screening for radiology staff, targeted resilience-building programs, guaranteed PPE supply, and awareness campaigns to reduce infection-related anxieties.
• The study aligns with broader literature showing health care workers’ mental health is influenced by PPE availability, exposure risk, symptom presence, and interpersonal emotional dynamics, with resilience acting as a buffer.

Limitations

• Geographic scope limited to Sichuan Province; may not generalize to other regions or countries.
• Sample included only radiology staff (nurses and technicians); findings may not apply to other departments with different workflows and stressors.
• Cross-sectional design; cannot establish causality or track anxiety changes over time.
• Self-reported measures; potential reporting bias and the lack of clinical diagnostic assessment.
• Environmental and family support factors not examined; unmeasured confounders could influence results.
• Timepoint captured early in the outbreak; anxiety trajectories may evolve as the situation develops.

Conclusions and practical takeaways

• At the start of the COVID-19 outbreak, anxiety among high-exposure radiology health care workers was high but mostly within normal limits, with a minority experiencing mild to severe anxiety.
• Key risk factors identified: older age, nurse role, lack of protective materials, presence of suspected symptoms in the participant, and higher susceptibility to others’ emotions.
• A strong protective factor identified: psychological resilience significantly associated with lower anxiety.
• Implications for practice:
  • Implement timely mental health assessments for radiology staff, especially those over 30, nurses, and those reporting PPE shortages or suspected symptoms.
  • Strengthen resilience through training and organizational support.
  • Ensure reliable PPE supply and clear infection control communications.
  • Consider symptom screening and rapid testing for staff with suspected symptoms to reduce anxiety and prevent further transmission.

Formulas and statistical references (LaTeX)

• SAS standard score calculation used in this study:
  $S = \left\lfloor 1.25 \times R \right\rfloor,$
  where $R$ is the rough score from the SAS questionnaire, giving a total score range of $S \in [25, 100]$.

• Anxiety severity categories based on SAS score:

  • No anxiety: S < 50
  • Mild anxiety: $50 \le S \le 59$
  • Moderate anxiety: $60 \le S \le 69$
  • Severe anxiety: $S \ge 70$

• Regression model statistics:
  R^2 = 0.251,\quad \text{Adj. } R^2 = 0.238,\quad F = 19.945,\quad P < 0.001.

• Multivariate linear regression coefficients (predictors of anxiety):

  • Presence of suspected symptoms in participant: B = -7.188,\ SE = 1.620,\ b = -0.208,\ t = -4.436,\ P < 0.001
  • Susceptibility to emotions and behaviours of people around them: $B = 2.804,\ SE = 1.027,\ b = 0.128,\ t = 2.731,\ P = 0.007$
  • Job function (nurse vs technician): $B = -2.085,\ SE = 0.902,\ b = -0.110,\ t = -2.311,\ P = 0.021$
  • Psychological resilience (low vs high): B = -8.454,\ SE = 1.114,\ b = -0.349,\ t = -7.590,\ P < 0.001
  • Availability of adequate protective materials: $B = -1.125,\ SE = 0.437,\ b = -0.122,\ t = -2.572,\ P = 0.011$
  • Age (≥30 vs <30): $B = 1.929,\ SE = 0.884,\ b = 0.105,\ t = 2.181,\ P = 0.030$

• Note on interpretation: the direction of some B-values reflects the coding of categorical variables in the regression; the observed group means indicate higher anxiety with symptoms, higher emotional susceptibility, and lower resilience, while adequate PPE and younger age were associated with lower anxiety in the multivariate context.

Key takeaways for exam preparation

• Recognize the main risk factors for anxiety among high-exposure radiology staff: age >30, nursing role, PPE shortages, presence of symptoms, and high susceptibility to others’ emotions.
• Recognize resilience as a critical protective factor against anxiety, reinforcing the value of resilience-building programs.
• Understand the importance of PPE supply and infection control confidence in reducing anxiety and maintaining work performance.
• Be able to interpret basic regression outputs: R^2, adjusted R^2, F-statistic, and standardized vs unstandardized coefficients (B, SE, b, t, P) in the context of psychological research.
• Appreciate limitations of cross-sectional, self-report studies and the need for longitudinal follow-up to track changes over time and the impact of interventions.