MIP (Norman, 2001; 2004) is a psychoanalytic method adapted to the requirements of the infant as analysand in the presence of his mother. In the study, the analysts strived to recruit the baby for an emotional interchange, though this did not imply any belief that the infant would understand verbal communication. Rather, the analyst addressed the baby to help him liberate emotions consolidated in symptoms such as screaming, avoiding maternal eye contact, and breast refusal. The analyst took care in enrolling the participant mother. This was to enhance her understanding of the baby's predicament and the nature of their relation, as well as giving her space to vent her own frustration, depression and anxiety.

The Symptom Check List-90 (SCL-90; Derogatis, 1994), with a Swedish language version (Fridell, Cesarec, Johansson, & Malling Thorsen, 2002), is a self-report questionnaire containing 90 items rated from 0 to 4. Higher scores indicate a worse outcome. The General Severity Index (GSI, or the mean across all items) was used to measure maternal general psychological distress.

Scheduled nurse calls at the local Child Health Centre (CHC), with paediatric checkups at 2 and 6 months of age. The nurse is encouraged to promote attachment and to detect postnatal depressions. Mothers may be offered parental groups, infant massage or guidance promoting interaction, as well as appointments with a paediatrician or a child psychiatric psychologist. Within the CHC framework, additional treatment may initiated by the nurse or the mother. This will be registered at the end-point interview.

Total of all reporting groups

Treatment as usual with nurse visits at Child Health Centres as part of regular Swedish child health care.

The EPDS (Swedish translation, Lundh & Gylland, 1990), is a self-report questionnaire containing 10 items each with a 3-point scale. Range: 0 - 30. Higher scores indicate a worse outcome. It is widely used at Swedish CHCs and has been validated on samples in Sweden.

Outliers (z-transformed scores>3.29) were replaced by raw scores corresponding to z=3.29. Multivariate outliers identified via Mahalanobi's distance through a multiple regression, none found. Missing data were very rare and missing at random. No scores imputed. Intention to treat (ITT)analysis was used.

Items are mostly rated on a 4-step scale, with 0,5,10 or 15 points per item, where 0 is most optimal. There are three versions for the age ranges of this study: 3-8, 9-14, and 15-20 months. To enable comparison across age groups we report mean scores across all items. Higher scores indicate a worse outcome. Each version was independently translated into Swedish, retranslated and approved by the constructor.

Outliers (z-transformed scores>3.29) were replaced by raw scores corresponding to z=3.29.Multivariate outliers identified via Mahalanobi's distance through a multiple regression, none found. Missing data were very rare and missing at random. No scores imputed. Intention to treat (ITT)analysis was used.

A Swedish-language version of the Parenting Stress Index (PSI; Abidin, 1990) with 35 items, each ranging 1-5 points. Higher scores indicate a worse outcome.

Outliers (z-transformed scores>3.29) were replaced by raw scores corresponding to z=3.29. Multivariate outliers identified via Mahalanobi's distance through a multiple regression, none found. Missing data were very rare and missing at random. No scores imputed. Intention to treat (ITT)analysis was used.

The EAS assessed video-taped mother-baby interactions of 10' duration on three maternal dimensions (Sensitivity, Structuring, Non-intrusiveness) and two infant dimensions (Responsiveness and Involvement. The raw scores of the subscales have different ranges (0-5, 0-7, and 0-9). To enable comparison across subscales, we divided scores in each subscale with its maximal score. This yielded a range for each subscale of 0-1.Thus, the total score range for all subscales was 0-1, with higher scores indicating a better outcome. Here we report results on Sensitivity.

Outliers (z-transformed scores>3.29) were replaced by raw scores corresponding to z=3.29. Multivariate outliers identified via Mahalanobi's distance through a multiple regression, none found. Missing data were very rare and missing at random. No scores imputed. Intention to treat (ITT)analysis was used.

The Symptom Check List-90 (SCL-90; Derogatis, 1994), with a Swedish language version (Fridell, Cesarec, Johansson, & Malling Thorsen, 2002), is a self-report questionnaire containing 90 items rated from 0 to 4. Higher scores indicate a worse outcome. The General Severity Index (GSI, or the mean across all items) was used to measure maternal general psychological distress.

Scheduled nurse calls at the local Child Health Centre (CHC), with paediatric checkups at 2 and 6 months of age. The nurse is encouraged to promote attachment and to detect postnatal depressions. Mothers may be offered parental groups, infant massage or guidance promoting interaction, as well as appointments with a paediatrician or a child psychiatric psychologist. Within the CHC framework, additional treatment may initiated by the nurse or the mother. This will be registered at the end-point interview.