How-To Guides

Higher mean desire to self-harm (DSH) was the clearest predictor of later NSSI in this cohort. Adults who engaged in NSSI during follow-up had higher DSH scores (mean = 53.96, SD = 33.65), and the mean DSH score on the day of NSSI was 59.00 (SD = 22.74). The discriminatory ability of mean DSH was acceptable, with an AUC of 0.743, whereas DSH variability was not associated with NSSI (Z = 0.408, P = .683; OR = 1.021, 95% CI = 0.925-1.127) and had poor discrimination (AUC = 0.563).

Passive suicidal ideation, not active suicidal ideation, was more closely associated with NSSI. When the investigators separated suicidal ideation into passive and active forms, they found a significant NSSI × ideation type interaction (F1,5097 = 40.74, P < .001). Patients who engaged in NSSI had higher passive SI than those who did not (mean ± SEM = 53.8 ± 3.90 vs 42.4 ± 1.68, P < .049), and passive SI increased across follow-up in the NSSI group (b [SE] = 0.029 [0.014], t = 2.30, P = .042).

Yes. The NSSI group had significantly higher negative affect during follow-up. Negative affect was higher in participants who engaged in NSSI than in those who did not (mean ± SEM = 64.5 ± 4.39 vs 54.8 ± 1.79; F1,94 = 4.22, P = .043). There was also a significant time × NSSI interaction (F1,2513 = 9.26, P = .002), and simple effects showed that negative affect decreased over time only in the NSSI group (b [SE] = -0.036 [0.01], t = -2.63, P = .009).

Restlessness was the specific emotion that was significantly higher in the NSSI group. In analyses of individual negative emotions, the interaction between NSSI and emotion type was significant (F8,2491 = 5.42, P < .001), but post hoc testing showed a difference only for restlessness. Restlessness was higher in the NSSI group than in the non-NSSI group (mean ± SEM = 69.6 ± 5.23 vs 45.1 ± 1.40, P = .005).

Overall interpersonal problems were not associated with NSSI in the mixed-model analysis, but lack of independence was higher in the NSSI group. The investigators did not find a significant overall association between NSSI and interpersonal problems, although interpersonal problems decreased over follow-up (b [SE] = -0.026 [0.01], t = -2.94, P = .003), with a larger decrease in the NSSI group (time × NSSI: F1,3129 = 11.06, P < .001; simple slope b [SE] = -0.055 [0.014], t = -3.89, P < .001). Among specific interpersonal experiences, lack of independence was higher in the NSSI group (mean ± SEM = 66.7 ± 5.71 vs 45.0 ± 2.66, P = .032).

No. This study did not find that intraindividual variability in suicidal ideation, negative affect, or interpersonal problems predicted NSSI. The authors specifically reported no statistically significant associations between NSSI and fluctuations in SI, negative affect, or interpersonal problems (all P values > .050). However, variability in SI was positively associated with the intraindividual mean of DSH (b [SE] = 0.481 [0.211], t = 2.28, P = .033).

This was a 12-month prospective ecological momentary assessment study in high-risk adults seen after a suicide attempt or severe suicidal ideation. The analysis included 106 adults recruited from psychiatric emergency and outpatient services in Madrid, Spain; 15 engaged in NSSI during follow-up. Participants used a smartphone app that sent 1 daily notification between 9:00 AM and 9:00 PM with 2-4 randomly selected questions from a 27-item pool, generating 12,408 observations in total; the mean number of responses per person was 72.14, and the average follow-up duration was 107.16 days.

The findings apply mainly to high-risk adults after a recent suicidal crisis, not to the broader NSSI population. Participants were adults aged 18 years or older who had presented within the previous month with a suicide attempt or a severe suicidal ideation episode requiring urgent care. The authors state that the sample was a specific cohort of high-risk adults post-crisis, so the results should not be generalized to adolescents, young adults, community samples, or broader NSSI populations without replication.

The authors suggest closely monitoring desire to self-harm, passive suicidal ideation, negative affect, and selected symptoms such as restlessness, anxiety, and lack of independence. In their conclusions, they state that DSH and passive SI were robust signals for predicting NSSI and that repeated monitoring through digital tools or structured follow-up may improve post-crisis risk assessment. They also note that restlessness, anxiety, and feelings of lack of independence were more prevalent in patients who engaged in NSSI during follow-up.

The main limitations were the small and homogeneous sample, the low number of NSSI cases, limited daily sampling, and the exploratory design. Only 15 participants reported NSSI, and the sample was culturally homogeneous and recruited based on suicidal ideation or prior suicide attempt rather than NSSI history. The app asked only 2-4 random questions once per day, which may have undersampled some variables and limited variability estimates; the authors also note that diagnoses and age were not controlled for and that the study was exploratory without prior hypotheses.