Participants

Sixty panellists answered the questionnaire at the first round, and 58 at the second round.

With regard to the geographic distribution, the panel was well representative of the Italian situation (33% North, 20% Centre, 47% South and Islands). Among the 60 participants 21 worked in universities, 39 in public structures as Department of Mental Health, organized in day-care Mental Health Centers or Hospital Diagnosis and Treatment of Psychiatric Services.

The overall result of the 17 statements is summarized in Table 1 and displayed in graphical form (level of agreement/disagreement distribution by box plot and bar graph) in Fig. 3.

Table 1 Summarization of the results of the 17 statementsFig. 3

Level of agreement/disagreement. Results distribution of each statement at Round 2

For 8 among the 17 statements of the Delphi more than 75% of panellists assigned a score of 4 or more reaching agreement (25th percentile ≥ 4). These statements also reached a high consensus (IQR < 2) as the panellists assigned similar scores. The 8 statements are:

8. When treating a depressed patient who has failed to respond to two antidepressants, I do tend to add lithium first.

For this statement, the IQR is 1.5 in the second round with a median value of agreement of 5% and 52% of respondents changed their answer between the first and second round. Only two cases assigned a score lower than 4 with the following motivations and one a score higher than 5.5:

I prefer augmentation with atypical antipsychotics rather than stabilizers: score 3

There are other possibilities as re-evaluation of possible organic causes, combination of antidepressant drugs, augmentation with atypical antipsychotics, esketamine: score 3

Based on my clinical experience, I obtained the most effective response using drugs: score 7.

10. When treating a depressed patient who has failed to respond to two antidepressants, I believe the best option is to combine esketamine (if currently available or when available in my centre).

For this statement the median value is 6, the IQR is 2 in both rounds and 27% of respondents changed their answer between the first and second round. Only one psychiatrist answered out of range with this motivation:

        •There are more accessible and cheaper enhancement strategies, as well as indicated, given the prevalence of bipolar depression: score 4.

11. I am satisfied with the efficacy of lithium and/or antipsychotics as augmentation therapies for patients with TRD.

For this statement, half of the panellists changed their minds between first and second round but at the end 100% of them agreed with the statement and assigned a score between 4 and 5. The IQR is 1 in both rounds. Three psychiatrists assigned a higher score with respect to the median declaring:

Based on my clinical experience, I believe that augmentation is very useful: score 6

I hardly do not get satisfactory results: score 7.

13. After obtaining a satisfactory response to treatment in a patient with TRD, long-term maintenance of therapy is essential.

For this statement at the second round, the 77% assigned score 7 (maximum agreement) and the rest of the group assigned score 6. Only 20% of panellists changed their answer, the lowest value of the entire survey, the IQR at the second round is 0, indication of the total agreement and consensus.

14. In my opinion, most patients with TRD can be treated with esketamine nasal spray in a community outpatient setting, without difficulties.

The median value of the agreement is 6, with values clearly in disagreement with the statement even at the second round. The IQR is 2 even at the second round, with 29% of panellists that varied from first and second round. It is of maximum interest to describe the motivations of the psychiatrists in disagreement with the statement at the second round:

Based on my experience, I believe that at least the induction phase should be administered in a hospital environment, possibly Day Hospital (DH), mainly due to the risk associated with dissociative phenomena: rate 2

The cost of the drug, the consequent difficulties in the governance of pharmaceutical expenditure and the need for the presence of personnel in the autonomous area make it difficult to think that it will be easy: rate 1

I agree but often the drug is not available: rate 4.

15. Educational support for patients helps to make the best use of the therapeutic opportunity offered by esketamine nasal spray.

At the second round, 100% of the participants reached agreement (score 6–7) and consensus, and the IQR is 1. The 27% varied their answers from first and second round.

16. In my daily reality, I have adequate and sufficient resources (staff, logistics, facilities, etc.) to provide patients with TRD with the best possible care.

The median value of agreement is 5. The IQR for this statement is wide: 3 in the first round, 2 in the second one; the 39% of respondents varied the answers. There are motivations in disagreement with the statement even at the second round:

The cost of the drug, the consequent difficulties in the governance of pharmaceutical expenditure and the need for the presence of personnel in the autonomous area make it difficult to think that it will be easy: rate 2

Unfortunately, the lack of doctors and inadequate facilities make it difficult to give adequate answers: rate 2.

17. In my opinion there are aspects of professional responsibility that the clinician must take into consideration in order to prefer, when possible, drugs with approved clinical indications for patients with TRD.

Almost all the participants assigned a score between 6 and 7 (maximum agreement); the IQR is 1 for both rounds. 32% of the participants varied their answers.

There are four statements classified as high consensus, but with indecision in the agreement (25th percentile < 4 and 75th percentile > 4 and IQR ≤ 2).

From a practical point of view, this means that the panellists assigned similar scores, but these scores are in the middle of the Likert scale showing a moderate agreement with the statement: less than 75% but more than 50% (in this case) assigned a score ≥ 4.

1. In my opinion, environmental and personological factors are among the main factors responsible for the non-response to treatment of almost all patients with TRD.

48% of the respondents varied their answers between first and second rounds.

4. In my opinion, a shared strategy is being pursued in Italy (based on guidelines, evidence-based treatments, diagnostic-therapeutic pathways) for the management of patients with TRD.

52% of the respondents varied their answers between first and second rounds.

5. Based on my clinical experience, I believe that after the second treatment failure there is a clear reduction in remission chances.

61% of the respondents varied their answers between first and second round (the widest variation recorded among the survey results). One member of the panel in the second round assigned a score out of the range that was reached in the first round, closely in agreement with the statement:

As reported in the literature, after the first two failures, the possibility of an improvement decreases drastically: score 7

For the following statement the panel reached the consensus with an IQR value of 2, but the mean value of 4.5 (3–5 at the second round) classified it as indecision. In particular, 36% of voters assigned a score of 3 and 47% voted 5 or 6.

9. In treating a depressed patient who has failed to respond to two antidepressants I do tend to associate psychotherapy first.

38% of the respondents varied their answers between first and second round. One psychiatry gave a motivation for his answer in disagreement with the statement, while two were more in agreement than the others:

It is often not feasible in a territorial context: score 2.

Because I believe that there may be personality factors that interfere with the response: score 7.

In my opinion, psychotherapy is essential in cases of resistance to treatment as very often the resistance is associated with important psychological or environmental factors: score 7.

The last group is represented by 5 statements classified as indecision in the agreement (the 25th percentile < 4, 75th percentile > 4) and low consensus with the statement (IQR ≥ 3).

From a practical point of view, this means that there is a low consensus among the panellists with score ranging from 2 to 7 in the Likert scale.

2. In my clinical practice, I also consider as treatment resistant a patient with incomplete improvement of symptoms (partial response) after an adequate period of treatment.

The panel did not reach the consensus around this statement with a IQR value of 3, and the answers values varying from 2 to 6 in the first round to 3–6 in the second round. 21% of the respondents assigned a score of 2 and 32% assigned a score of 6. 38% of the respondents varied their answers between the first and the second rounds.

3. In clinical practice, I usually use scales/questionnaires for the diagnostic classification and/or evaluation of the clinical course of the patient with depression.

The panel did not reach the consensus around this statement with a IQR value of 3, with a huge variability in the scores of the answers at the second round (from 2 to 7). 30% of the respondents assigned a score of 3 and the 39% assigned a score of 6. 50% of the respondents varied their answers between the first and the second rounds.

Those who answered out of the range of the first round are the following:

I do not use them in current practice, mainly seeing patients in the private practice: score 2.

It is often not feasible, given time constraints. The electronic record systems provided by the public service are not adequate for the integration of the psychometric evaluation: score 2.

Fully agreement because I work in a university and research setting: score 7.

6. When I treat a depressed patient, if after 3–4 weeks there is no response, I decide to change the antidepressant therapy.

The panel did not reach the consensus around this statement with a IQR value of 3 in both rounds, with a mean value at the second round of 4.5 (indecision). 55% of the respondents varied their answers between the first and the second round.

At the second round 1 psychiatrist gave a motivation completely in disagreement with the statement, while on the other hand 2 colleagues gave motivations fully in agreement:

Considering the latency of response to oral serotonergic drugs (at least 3–4 weeks) before changing the dosage or the antidepressant or boosting with lithium or other drugs, I usually wait 6–7 weeks from the start of treatment: score 1.

I believe that it is correct to change therapy because, according to my experience, it is unlikely that after this period if there has been no clinical improvement, this will occur later: score 7.

I believe that it is essential to evaluate early signs of response, including side effects, as predictors of an effective response at 2 months: score 7.

7. When treating a depressed patient who has failed to respond to two antidepressants, I believe that the best option is to combine an antipsychotic, such as quetiapine.

The panel did not reach the consensus around this statement with a IQR value of 3 in both rounds, with a mean value at the second round of 3, with the first and third quartile of 3 and 5 respectively the level of agreement is classified as indecision. At the second round the answers are spread on all 7 possible scores, about 20% for each of the possible scores between 2 and 5. 43% of the respondents varied their answers between the first and second rounds. Two motivations strongly in disagreement and three closely in agreement with the statement are recorded:

I prefer to use aripiprazole: score 1.

I rarely choose quetiapine for augmentation in these cases: score 1.

Trying an antipsychotic is a valid augmentation strategy: score 6.

It depends on the specific psychopathological dimensions of the patient: score 6.

Several evidence have shown that it is not advantageous in terms of efficacy to further switch antidepressants, but rather to use augmentation or combination strategies: score 7.

12. I use (or I will use) esketamine nasal spray only after non-response to the available augmentation/increase strategies.

The panel did not reach the consensus around this statement with a IQR value of 2.5 at the second round, with a mean value of 5, with the first and third quartiles of 3 and 5.5, respectively; the level of agreement is classified as indecision. 48% of the respondents varied their answers between the first and the second rounds.