Lieberman 1991.

Methods	Randomised, 2‐arm comparison PIP of anxiously attached 12‐month‐old infants versus control Unit of allocation: participant Study dates: not reported
Participants	Number randomised: n = 59 dyads randomised Number randomised to each group: intervention n = 34; control n = 25. Securely attached dyads formed a second non‐randomised control group (n = 34), not addressed in this review Participants: mother and infants (dyads) received therapy Sex of infants: intervention 52.9% boys; control 44% boys Age of child: 12 months (11‐14 months at entry) Maximum age of child at entry: up to 14 months Ethnicity: Hispanic Recruitment: the dyads were recruited from paediatric clinics at a large teaching hospital and neighbourhood health clinics Risk factors: at risk for disorders of attachment because quote: "recent Latino immigrants face a high incidence of depression and anxiety as a result of poverty, unemployment, and cultural uprootedness". "We surmised that the infants would be at risk for anxious attachment because their mothers might be relatively emotionally unavailable due to their circumstances" (p. 200) Social and economic status: economic hardship was reflected in the living conditions and parents' employment with 74.2% sharing living quarters and only 44.3% fathers and 8.2% mothers working full time in low‐paid service jobs. Most families shared living quarters with others (74.2%). Only 44.3% of fathers and 8.2% of mothers worked full‐time, mostly in low‐paying service jobs Country and Setting: USA; intervention delivered during home visits with laboratory outcome assessments Child development: no congenital birth defects; overall 29.6% of whole sample had some birth complication; assumed to be typically developing Inclusion criteria: children aged 12‐24 months at entry; at risk due to stressful life events. At 12 months of age, during a home visit, infants were assessed using the SSP (Ainsworth 1978). Infants classified as anxious‐avoidant or anxious‐resistant were randomised Exclusion criteria: > 5 children to each mother; flagrant psychological disorders Comparability of randomised groups at baseline: no differences except in unemployment rates: 15.0% anxious control; 53.8% anxious intervention; P value = 0.024
Interventions	Duration of intervention: 12 months Number of sessions and frequency: weekly sessions. Total number of sessions not stated. The study began when the infants were 12 months old and ended shortly after their second birthday Intervention: quote: "The intervention began immediately after research group assignment. The format was unstructured weekly sessions with mother and baby. Visits took place mostly in the home and lasted 1 1/2 hours. Each dyad had one intervenor during the entire intervention period. The main focus of the intervention was to respond to the affective experiences of mother and child, both as reported by the mother and as observed through the mother‐child interaction. There was no didactic teaching. Instead, the intervenors sought to alleviate the mothers' psychological conflicts about their children and to provide developmental information that was clinically timed and tailored to the child's temperament and individual style. The developmental information focused on areas relevant to quality of attachment, such as contingency to signals, availability of age‐appropriate opportunities for exploration, and negotiation of infant‐mother conflicts to promote a goal‐corrected partnership. In attachment theory language, this intervention approach intends to provide the mother with a corrective attachment experience. The intervenor spoke for the mother's affective experience, addressing the legitimacy of her longings for protection and safety both when she was a child and currently as an adult, and enabling her to explore unsettling feelings of anger and ambivalence toward others (including the child and the intervenor). The intervenor linked this process to the child through appropriately timed developmental information to reduce negative attributions and to support a benign perception of the child's motives. When feasible, the intervenor also tried to provide concrete elements of protection by helping to secure needed goods and services" (p. 201) Type of psychodynamic model: based on Fraiberg 1980 Therapist: the 4 intervenors were bicultural, bilingual women with Master's degrees in psychology or social work and with clinical experience Control: not stated, assumed to be TAU Monitoring fidelity of the intervention: each intervenor received weekly supervision by senior faculty
Outcomes	Timing of outcome assessment: post‐intervention (24 months of age) Outcomes Parent: LEI (stressful life events) (Egeland 1979) at baseline Parent/infant interaction: free play measures; maternal empathic responsiveness/empathy (equivalent to the Home Observation for Measurement of the Environment (HOME) Inventory: maternal care, maternal responsiveness, and maternal involvement Bradley 1977); maternal initiation of interaction (7‐point scale); behaviour on reunion (goal‐corrected partnership, 9‐point scale); maternal child‐rearing attitudes (control aggression, encourage reciprocity, complexity) (Cohler 1970) Child: restriction of affect (7‐point scale); angry behaviour (frequency); security of attachment (Q‐sort) (Waters 1985). Avoidance and resistance, behaviour on reunion, and goal‐corrected partnership were measured using Ainsworth Interactive Behaviour Scales, which are modelled after the SSP, but with longer 10‐minute episodes suitable for younger children Ainsworth 1978) Adverse effects: none Clinician or participant report of outcomes: all outcomes reported objectively by clinician/investigators
Notes	Funder: NIMH Prevention Research Branch grant 5 ROI MH39973 to AF Lieberman Aim: to test the hypothesis that PIP can improve quality of attachment and social‐emotional functioning Study investigator's conclusion: intervention group infants were significantly lower than anxious controls in avoidance, resistance, and anger. They were significantly higher than anxious controls in partnership with mother. Intervention mothers had higher scores than anxious controls in empathy and interactiveness with their children. The groups did not differ in maternal child‐rearing attitudes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Anxiously attached dyads were assigned to the intervention or the control group by block randomization for infant sex and birth order (first, later). Fifty‐nine dyads were randomly assigned (intervention group n = 34 and anxious control group n = 25)" (p. 201)
Allocation concealment (selection bias)	Low risk	Further details from investigator: allocation of participants was concealed up to the point where treatment began, the randomisation took place during the last assessment session after the SSP classification took place. At that time, the clinician opened an envelope in the mother's presence where a piece of paper stated which group the dyad would be assigned to: intervention or control. The clinician found out at the same time as the mother about the group assignment, which was done at random (email 3 June 2013)
Incomplete outcome data (attrition bias) All outcomes	Low risk	32/59 completed. 7 cases dropped out after the home visit but before the pre‐intervention SSP (12 months of age) and could not be assigned. Overall attrition was 18% and did not differ for the experimental groups. At the post‐intervention assessment (24 months of age), the randomised sample consisted of 82 dyads: 29 in the anxious‐intervention group and 23 in the anxious‐control group
Selective reporting (reporting bias)	Low risk	All pre‐specified outcomes are reported, but the LEI (parental stress) result at post‐intervention was presented only as a whole sample mean
Blinding of participants and personnel (performance bias) All outcomes	High risk	No details given. Judged as high risk after we obtained further information from the investigator as it is likely that given the nature of the intervention the participants or personnel would have been aware of their assignment to therapy and, at randomisation, the mothers were told if they had been assigned to intervention or control
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The Strange Situation was coded by fully trained and experienced judges. The home and laboratory measures were scored by coders who were blind to the 12‐month classification and group assignment of the dyads. Different coders scored the home and laboratory measures. In the 24‐month laboratory situation, the free‐play measures and reunion measures were scored by different coders. Intervention process raters had no access to other outcome data. These procedures were adopted to protect the independence of the different ratings" (p. 203)
Other bias	Low risk	No other sources of bias; no imbalance in terms of participant characteristics at baseline