When zero isn't nothing: excess zero scores in repeated distress screening of cancer patients

This study examined patterns of psychological distress among a large sample of oncology outpatients and evaluated the validity of repeated ultra-brief screening measures given unexpectedly high rates of zero scores.

Distress was assessed using the distress thermometer (DT) and Patient Health Questionnaire-2 (PHQ-2) administered routinely in outpatient cancer care between July 2021 and July 2023. Data were extracted from the electronic health record (EHR) and analyzed using zero-inflated negative binomial (ZINB) models to estimate distress severity and the likelihood of "excess zero" scores beyond those expected from the distribution. Covariates included sex, age, race, ethnicity, cancer type, disease stage, time since diagnosis, and screening frequency.

Across 39,762 screening instances from 7081 patients, most scores were zero (PHQ-2 = 92%; DT = 91%). ZINB models indicated substantial zero inflation (PHQ-2 = 59%; DT = 90%). Female sex, Black racial identity, and advanced cancer were associated with higher distress severity (ps < 0.01), while the probability of excess zero scores increased with both time since diagnosis and number of repeated screens (ps < 0.001). A prior positive depression screen was associated with a one-third reduction in the odds of subsequent excess zeroes (p < 0.001).

Findings revealed strikingly high rates of zero distress scores, suggesting potential underreporting or measurement limitations in repeated ultra-brief screening. Implementation factors such as screening frequency, mode, and patient context may contribute to apparent "false zeros." Programs should consider complementary assessment methods and optimized intervals to more accurately identify patients in need of psychosocial support.