3 results
A natural language processing approach to modelling treatment alliance in psychotherapy transcripts
- Jihan Ryu, Stephen Heisig, Caroline McLaughlin, Rebeccah Bortz, Michael Katz, Xiaosi Gu
-
- Journal:
- BJPsych Open / Volume 7 / Issue S1 / June 2021
- Published online by Cambridge University Press:
- 18 June 2021, p. S48
-
- Article
-
- You have access Access
- Open access
- Export citation
-
Aims
Patient-therapist alliance is a critical factor in psychotherapy treatment outcomes. This pilot will identify language concepts in psychotherapy transcripts correlating with the valence of treatment alliance using natural language processing tools. Specifically, high-order linguistic features will be extracted through exploratory analysis of texts and interpreted for their power to discriminate alliance rated by patients.
MethodAdult patients and therapists in outpatient clinic at various stages of relationship building and treatment goals consented to participate in the cross-sectional study approved by the Institutional Board Review. Psychotherapy sessions were recorded using wireless microphones and transcribed by two research assistants. After the recording, each patient completed Working Alliance Inventory– Short Form, to generate clinical scores of alliance. We used the Linguistic Inquiry Word Count (LIWC) tool to map words to psycholinguistic categories, and generated novel linguistic parameters describing the individual language for each speaker role. Canonical-correlational analysis and descriptive statistics were used to analyze the two datasets.
ResultPatients (N = 12, 83% female, mean age = 40) were primarily diagnosed with personality disorders (67%) working on real-life interpersonal issues (median treatment duration 18.5 weeks, 50% psychodynamic, 32% cognitive-behavioral, 16% supportive modality). In this heterogenous sample, patients who used the “achieve” (e.g. trying, better, success, failure) and “swear” psycholinguistic categories of words rated the treatment alliance lower (r=−0.70, p = 0.01; r=−0.65, p = 0.02). Patients rated alliance lower with therapists, who used more “I” pronoun (r=−0.58, p < 0.05) and higher with therapists using more “risk” (difficult, safe, crisis) and “power” (important, strong, inferior, passive) categories (r = 0.66, p = 0.02, r = 0.58, p < 0.05), which commonly appeared in psychoeducation and conceptual framing of problems. Interestingly, there was no correlation with “affiliation” category (p = 0.9). Linear regression modeling from “achieve,” “swear” variables and “I,” “risk” variables with duration of treatment as covariate predicted the patient's rating of alliance (Adjusted R2 = 0.66, p = 0.03).
ConclusionOur data collection and sub-sample analysis are ongoing. Preliminary results are showing speaker-specific language patterns in cognitive-emotional domain, e.g. self-expressivity, and in clinician's therapy style, covarying with the patient's perceived closeness in the heterogenous treatment dyads. Novel application of natural language processing to characterize alliance using the data-driven approach is an unbiased method that can provide feedback to clinicians and patients. This characterization can also potentially provide insights into the mechanisms underlying the therapeutic process and help develop psycholinguistic markers for this critical clinical phenomena.
Effect of folic acid supplementation during pregnancy on brain health of the child at 11 years: the FASSTT Offspring trial
- Aoife Caffrey, Kristina Pentieva, Helene McNulty, Pramod Gaur, Joel B. Talcott, Caroline Witton, Anthony Cassidy, Marian McLaughlin, Diane J. Lees-Murdock, Rachelle E. Irwin, Colum P. Walsh, Girijesh Prasad
-
- Journal:
- Proceedings of the Nutrition Society / Volume 79 / Issue OCE2 / 2020
- Published online by Cambridge University Press:
- 10 June 2020, E507
-
- Article
-
- You have access Access
- Export citation
-
Periconceptional folic acid (FA) has an established role in the prevention of neural tube defects (NTDs), leading to global recommendations for FA supplementation before and in early pregnancy. However, it is unclear whether there are any benefits for offspring brain health arising from continued maternal FA supplementation beyond the first trimester. The aim of this study was to investigate the role of maternal folate nutrition during pregnancy in relation to cognitive performance and brain function in the offspring at 11 years. The children of mothers who had participated in a randomised trial of Folic Acid Supplementation in the Second and Third Trimesters (FASSTT) were investigated, providing a unique opportunity to examine offspring brain health in relation to maternal folate (the FASSTT Offspring trial; n = 68). Cognitive performance was assessed using the Wechsler Intelligence Scale for Children, Fourth UK Edition (WISC-IV). The WISC-IV measures Full Scale IQ and specific domains of cognitive performance: Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed. Brain function was measured using magnetoencephalography (MEG) in a subset of the child participants (n = 33). The results showed no significant difference in Full Scale IQ between the children of mothers who had received folic acid versus placebo during pregnancy (P = 0.993). Processing Speed subtest scores were however significantly higher in the folic acid group compared with placebo (Symbol Search: P = 0.046 and Cancellation: P = 0.011). The application of MEG analysis showed that at rest, there were differences in brain functioning with significantly lower overall power at Broad band [1–48Hz] (P = 0.041) and a trend (not significant) towards lower power in all other frequency bands (Theta, Mu, Beta, Low Gamma and High Gamma) in children from the FA group compared with placebo. Results for the responses to the language task (congruent and incongruent sentences) in children from the FA group showed significantly lower power within the Theta band [4–8Hz] and significantly higher power within high frequency bands i.e. Beta [13–30 Hz] and High Gamma [49–70 Hz]. This suggested more efficient language processing abilities in these children compared to children of mothers in the placebo group. The findings provide scientific evidence that continuing FA supplementation beyond the periconceptional period that is protective against NTDs, may be beneficial for brain health in the offspring. Furthermore, this study demonstrates that MEG is a useful tool for objective assessment of functional brain activity in healthy children in response to nutrition intervention.
Paying for Long-Term Domiciliary Care: A Comparative Perspective
- CAROLINE GLENDINNING, MICHAELA SCHUNK, EITHNE McLAUGHLIN
-
- Journal:
- Ageing & Society / Volume 17 / Issue 2 / March 1997
- Published online by Cambridge University Press:
- 01 March 1997, pp. 123-140
- Print publication:
- March 1997
-
- Article
- Export citation
-
Concerns over growing numbers and proportions of older people in industrialised societies have prompted interest in the development of cheaper ways of providing long-term care for older people. While debate in the UK is currently focused on the costs of residential and nursing care, other European and Nordic countries have introduced schemes designed to encourage or sustain the provision of ‘social’ care by family members, friends and ‘volunteers’, on the assumption that this can be provided at lower net public expense than either residential care or formally-organised domiciliary services.
Drawing on material from a detailed comparative study, this paper describes four different models on which such payments are currently based. These models are discussed and evaluated, taking into account factors which include the eligibility criteria for payments; maximising the autonomy of older people and family care-givers; and the relationships between financial payments and access to services.
These models locate systems of payment within the broader context of financial and service support designed to help frail older people and those who support them. They therefore highlight the importance of considering both financial support and services in comparative studies of social welfare provision. However, further evaluation and policy development is hindered by the lack of evaluation of different models of paying for care and a lack of evidence about the experiences of older people and care-givers.