Missing Is Not Yearning: A Russian Study Sorts Grief by How the Person Died
- Interview study of 105 bereaved Russian-speaking adults aged 20 to 41 (M = 28.3, SD = 7.69), 90 women (85.7%) and 15 men (14.3%), one to eleven years after the death (M = 5.03, SD = 3.59), split by cause: illness 48 (45.7%), old age 24 (22.9%), accident or violent death 33 (31.4%).
- Intensity of missing differed significantly by cause of death (Kruskal-Wallis H = 11.043, p = 0.004): lowest mean rank after death from old age (40.44), highest after accidental death (65.82), illness in between.
- Reported change in intensity over time also differed (chi-square = 18.678, p < 0.001). "It got easier" predominated after illness and old age, while the accidental-death group more often reported no change at all (chi-square = 12.223, p = 0.002).
- Across all three groups a common core held: sadness dominant, memories and mental addressing of the deceased, trigger-bound onset, and ambivalence of painful and positive content. Intercoder agreement (Cohen's kappa) was 0.82; a category counted as core above 40% of mentions.
ICD-11 put persistent yearning at the centre of prolonged grief disorder, which means every clinician now has to judge, in the room, whether the longing in front of them is the diagnostic core or the ordinary work of missing someone. Russian clinical vocabulary makes that judgement harder than it needs to be: "тоска" (yearning) and "скучание" (missing) are used almost interchangeably. A group at Kostroma State University tried to separate them empirically, and found the separation tracks how the person died.
What the Kostroma data show
The design is a cross-sectional interview study, not a cohort. One hundred and five adults, recruited through social-media announcements and interviewed individually, aged 20 to 41 (M = 28.3, SD = 7.69), 85.7% women. Inclusion required both a bereavement and a self-reported current experience of missing the deceased; where there were several losses, respondents chose the one that mattered most. Time since death ran from one to eleven years (M = 5.03, SD = 3.59). In the illness group the leading causes were cancer (25%), cardiovascular disease (18.6%) and COVID-19 (12.5%). Coding used interpretative phenomenological and content analysis; categories were then compared with Pearson chi-square and Kruskal-Wallis H in SPSS 22.0.
The emotional texture separated cleanly. Sadness was reported more often after death from illness (chi-square = 10.625, p = 0.005), as were emptiness (chi-square = 10.063, p = 0.007), a felt lack (chi-square = 16.089, p < 0.001) and a need for contact and support (chi-square = 10.063, p = 0.007). Regret predominated after old age and accident (chi-square = 13.84, p = 0.001). The form the experience took split three ways (chi-square = 15.646, p = 0.004): acute episodic bouts after illness, a background or situationally cued state after old age and after accident.
The time course is the clinically loaded part. The accidental-death group more often reported bodily reactions such as stupor and goosebumps (chi-square = 13.884, p = 0.001), a sense of having been personally changed (chi-square = 7.841, p = 0.02), and missing triggered by memories (chi-square = 6.397, p = 0.041) and by being alone (chi-square = 6.469, p = 0.039). The old-age group instead emphasised the value of life and relationships (chi-square = 21.69, p < 0.001) and named memory of the person as a positive aspect (chi-square = 8.701, p = 0.013); absence of bodily reaction was also most common there (chi-square = 8.092, p = 0.017). The authors read this as three patterns: wave-like after illness, background after old age, disorganising after sudden death.
What this changes at the table
Two things are usable immediately. First, ask how the person died before you interpret the longing. In this sample the same reported symptom carried a different trajectory depending on cause, and the group with the flattest trajectory was not the group with the most obviously dramatic presentation at intake but the one that had lost someone suddenly. Non-decay of intensity is a more informative signal than intensity itself, and it is a question you can ask directly rather than infer from a score.
Second, the invariant core is a useful reassurance to hand a patient. Sadness, recurring memories, silently addressing the deceased, sudden onset at a trigger, and holding pain and warmth in the same breath appeared across all three groups, including the ones adapting well. Within a continuing-bonds frame these are not markers of a stalled process. What the authors treat as the pathological pole is narrower: an intense, disorganising state focused on existential emptiness rather than on the person, with no adaptive yield.
The caution runs the other way too. No prolonged-grief instrument was administered here. Nobody in this study was diagnosed, and cause of death is a prompt for a better conversation, not a risk score. A convenience sample of mostly young women interviewed once cannot tell you the prevalence of anything.
Intensity that fails to decay is a more useful clinical signal than intensity that is merely high.
This is a convenience sample recruited via social media, 85.7% women, aged only 20 to 41, with unequal groups and no standardised prolonged-grief measure, so the missing/yearning boundary described here is phenomenological rather than diagnostic. Change in intensity over time was recalled retrospectively in a single interview, not measured longitudinally, and self-reported trajectories are vulnerable to current mood.