Characteristics of Included Studies. The majority of the studies were conducted in the USA (n = 8) with one study conducted in England. Five used a wider area which included participants from more rural settings, and four related to a single city and were restricted to urban areas. Most of the studies used adult samples (n = 8) and one used a sample of juvenile offenders. Of the studies in the USA, two recruited samples from jails, three from state prisons and three from mixed correctional facilities. The prison system operates in a different way and the study in England recruited from remand prisons. None of the studies were restricted to a single disorder and criteria for inclusion in the studies ranged from solely being diagnosed with a mental health problem, being treated by a mental health team within the prison, being adjudged to be of high risk, or being homeless before entry into custody. Four studies were based on cohorts, either from facilities that did not offer the intervention or from a time when the intervention was not available in the same facility. Two studies were randomised controlled trials. Two were case series with no comparison group and one used a pre-post comparison with outcomes compared to the same time period before contact with the program. Study outcomes ranged from contact with health services (n = 2), Medicaid enrolment (n = 1), recidivism (n = 4), sanctions for treatment non-compliance (n = 1) and place of residence at time of treatment discharge. Most of the studies were bridging interventions, with intervention provided both before and after release, but there were also examples of care only in the pre (n = 1) or post (n = 2) release period. The majority of the interventions used a mixed approach (n = 8) with case management, psychosocial modules and onward referral represented. The other study focused on Medicaid enrolment (n = 1). In the majority of cases, the intervention was delivered by a clinician (n = 7) and in two the required qualifications were not reported. Table 1 gives a more detailed description of the interventions used in each study. Records identified through database searching (n = 12044) Additional records identified through other sources (n = 33) Full-text articles assessed for eligibility (n = 46 ) Articles included (n = 10) Unique studies included (n = 9) Included Eligibility Full-text articles excluded, (n = 36) - Conference proceeding (6) - Not data based (4) - No intervention (7) - Not aimed at transition (7) - Not prison (4) - Not m...
Characteristics of Included Studies. Presented in Table 3 is the summary of the main characteristics of each study included in the systematic review and meta-analyses. The *next to the authors name denotes studies that were included in the meta-analysis. The subscript a next to the authors name signifies that studies contained the same HC comparison group. The column headed ‘Outcome of Measure’ gives a description of the comparison of the ES between clinical group and control group. The column titled ‘Group and N’ shows which patient group (e.g. AN, BN, ED) is being compared to the control group. For example the study by Chassler (1997) compares attachment between ED and controls. The table does not give a comparison between two clinical groups. Self-report instruments were used mainly to assess the areas of ‘Affiliation and attachment’, Self- Evaluation’ and ‘Social Dominance’, while experimental tasks were primarily used to test the areas of ‘Reception of Facial Communication’, ‘Production of Facial Communication’, ‘Reception of Non-Facial Communication’, ‘Agency’ and ‘Understanding Mental States’. Only in the area of ‘Production of Non-Facial Communication’, was there the same number of self-report and experimental tasks. The majority of studies investigated a sample of individuals with a current diagnosis of AN. Table 3. Study characteristics of studies included in the review of social processing in eating disorders Studies Group and N Age: mean (sd) BMI (sd) Co-morbidity Measure Outcome of measure Xxxxx’x d (95% C.I.) ‘Affiliation and Attachment’ *Amianto et al. (2011) AN 38 26.31 (7.4) 16.42 (2.0) NR Attachment Style Questionnaire Parental Bonding Instrument Insecurity Lower security Lower Parental Care Higher Parental Overprotection 0.779 (0.335, 1.223) 0.981 (0.534, 1.427) 0.900 (0.457, 1.343) 1.229 (0.768, 1.689) HC 50 23.95 (2.1) 21.86 (2.7) - *Bonne (2003) BN 16 23 (2) 19.96 NR Parental Bonding Instrument Lower Parental Care Higher Parental Overprotection 0.803 (0.081, 1.525) 0.254 (-0.441, 0.950) HC 16 23 (2) 19.96 - *Bulik et al. (2000) AN 15 AN-Rec 21 23-45 18.5 (2.6) NR Parental Bonding Instrument Lower Parental Care Higher Parental Overprotection 1.123 (0.560, 1.686) 0.408 (-0.066, 0.883) 0.379 (-0.167, 0.925) 0.496 (0.020, 0.972) HC 98 23-45 25.6 (6.5) - *Calam et al. (2006) AN 31 BN 67 25 - NR Parental Bonding Instrument Lower Parental Care Higher Parental Overprotection 0.104 (-0.270, 0.478) 0.438 (0.166, 0.711) 0.266 (-0.108, 0.641) 0.190 (-0.081, 0.461) HC 242 22.9 - - *X...
Characteristics of Included Studies. TRIPOD statement: a preliminary pre-post analysis of reporting and methods of prediction models The PubMed search string retrieved 481 articles, of which the full-text of 119 were read and 70 met our inclusion criteria (pre-TRIPOD: 32 articles, post-TRIPOD: 38 articles, Figure 1, Supplementary Text 1). Most of the included articles were published in The BMJ (n=38), and least in The Lancet (n=3) and NEJM (n=1). In both the pre- and post-TRIPOD period the majority of articles described prognostic models (as opposed to diagnostic models) and this increased in the post-TRIPOD period (pre-TRIPOD: 59%, post-TRIPOD: 89%) (Table 2). In the post-TRIPOD period the percentage of articles describing both the development and validation of a model (pre-TRIPOD: 31%, post-TRIPOD: 39%) or solely the external validation (pre-TRIPOD: 13%, post-TRIPOD: 26%) increased too. Thirty-two percent of articles only described the development of a prediction model without external validation in the post-TRIPOD period, compared to 44% in the pre-TRIPOD period. Figure 1: Flow chart of search results and selection procedure 185 Table 2: Characteristics of included studies Before 2015 (n=32) number, (%) After 2015 (n=38) number, (%) Diagnostic/Prognostic Diagnostic 13 (41%) 4 (11%) Prognostic 19 (59%) 34 (89%) Type Development 14 (44%) 12 (32%) Validation 4 (13%) 10 (26%) Development and Validation 10 (31%) 15 (39%) Update 4 (13%) 1 (3%) Setting General population and Primary care 18 (56%) 18 (47%) Secondary care 14 (44%) 20 (53%) Design Cohort 26 (81%) 31 (82%) RCT 1 (30%) 4 (11%) Cohort and RCT 2 (6%) 3 (8%) Case-Control 3 (9%) 0 (0%) Topic (Cardio)vascular 12 (38%) 16 (42%) Oncological 3 (9%) 8 (21%) Other 17 (53%) 14 (37%) Chapter 7 186 The majority of models were developed and/or validated using data from observational cohorts (pre-TRIPOD: 81%, post-TRIPOD: 82%) compared to other study designs such as randomised trials. More than half of the articles published in the post-TRIPOD period referred to the TRIPOD Statement (n=20, 53%) and were published in journals that published the TRIPOD Statement (n=21, 55%). The TRIPOD Statement was cited in 48% of articles published in journals that published the TRIPOD, and in 59% of articles in journals that did not publish TRIPOD.
Characteristics of Included Studies. Among the included studies, all were published on or after 2005, and analyzed data collected as early as 1991 (Table 2). Most of the studies were secondary analyses of Demographic and Health Surveys (DHS) (n=11) conducted in urban and rural settings in lower-middle income countries (n=8) and representative of two major geographical regions: Africa (n=4) and South-Asia (n=7). In addition, one African multi-country review was included. India was the most frequently represented country in the review (n=5), followed by Nigeria (n=3). The sample sizes ranged from around 1,000 to approximately 15,000 respondents. The most common method of analysis used by the authors was multivariate logistic regression (42%), with the remainder employing multilevel, stepwise or bivariate regression analysis. Only two of the studies included a theoretical framework that addressed both empowerment and immunization coverage. The most common age range for reporting immunization outcomes was 12-23 months (n=6, Table 2). Exposure Measure In general, the authors defined agency in diverse ways, most frequently “autonomy” (50%) and “empowerment” (17%). The items used to measure agency fell into three domains: decision-making, freedom of movement, and other (Table 3). Several reoccurring themes persisted across these items: health decisions of the woman, financial control, permission to go/ visit health centers, friends and family as well as large and daily household purchases. Across all studies, the most common measures of women’s agency were measures based on decision-making items (58%, Table 2). These studies used a total of 13 different DM items (Table 3), only two used DM items pertaining to a woman’s ability to decide to go to a health facility: “can go to the local health center without seeking permission”(DM 2, Table 3) and “decisions to go to a health facility” (DM 8, Table 3). Moreover, only one item specifically connected the mother to the child’s health by asking who was “the main decision-maker when the child is ill” (DM 1, Table, 3). Otherwise, the remaining DM items pertained to decisions regarding household purchases, ability to move and see friends, family, or relatives, finances, and health (Table 3). Less emphasis was placed on measuring agency with freedom-of-movement items only (without measuring decision-making) (n=2, Table 3). Both of the studies that included freedom-of- movement items analyzed data from India [17,20]. Many of the authors operationalized freedom ...
Characteristics of Included Studies
