Automated data analysis of unstructured grey literature in health research: A mapping review

The amount of grey literature and 'softer' intelligence from social media or websites is vast. Given the long lead-times of producing high-quality peer-reviewed health information, this is causing a demand for new ways to provide prompt input for secondary research. To our knowledge, this is the first review of automated data extraction methods or tools for health-related grey literature and soft data, with a focus on (semi)automating horizon scans, health technology assessments (HTA), evidence maps, or other literature reviews. We searched six databases to cover both health- and computer-science literature. After deduplication, 10% of the search results were screened by two reviewers, the remainder was single-screened up to an estimated 95% sensitivity; screening was stopped early after screening an additional 1000 results with no new includes. All full texts were retrieved, screened, and extracted by a single reviewer and 10% were checked in duplicate. We included 84 papers covering automation for health-related social media, internet fora, news, patents, government agencies and charities, or trial registers. From each paper, we extracted data about important functionalities for users of the tool or method; information about the level of support and reliability; and about practical challenges and research gaps. Poor availability of code, data, and usable tools leads to low transparency regarding performance and duplication of work. Financial implications, scalability, integration into downstream workflows, and meaningful evaluations should be carefully planned before starting to develop a tool, given the vast amounts of data and opportunities those tools offer to expedite research.

Evaluating high-cost technologies - no need to throw the baby out with the bathwater

INTRODUCTION

Evidence generation for the health technology assessment (HTA) of a new technology is a long and expensive process with no guarantees that the health technology will be adopted and implemented into a health-care system. This would suggest that there is a greater risk of failure for a company developing a high-cost technology and therefore incentives (such as increasing the funding available for research or additional market exclusivity) may be needed to encourage development of such technologies as has been seen with many high-cost orphan drugs.

AREAS COVERED

This paper discusses some of the key issues relating to the evaluation of high-cost technologies through the use of existing HTA processes and what the challenges will be going forward.

EXPERT OPINION

We propose that while the current HTA process is robust, its evolution into accommodating the incorporation of real-world data and evidence alongside a life-cycle HTA approach should better enable developers to produce the evidence required on effectiveness and cost-effectiveness. This should lead to reduced decision uncertainty for HTA agencies to make adoption decisions in a more timely and efficient manner. Furthermore, budget impact analysis remains important in understanding the actual financial impact on health-care systems and budgets outside of the cost-effectiveness framework used to aid decision-making.

The association between maternal and paternal substance use and child substance use, internalizing and externalizing problems: a systematic review and meta-analysis

AIMS

There is substantial evidence showing an association between parental substance use and child substance use and/or mental health problems. Most research focuses upon maternal substance use, with the influence of paternal substance use often being overlooked. We aimed to investigate the differential effects of maternal and paternal substance use upon children aged 0-18 years.

METHODS

We used systematic review methods to identify observational studies examining the association between either maternal or paternal substance use and child substance use and/or mental health problems. The odds ratio (OR) effect measure was used, for ease of computation. We used a random-effects model with the inverse variance method to meta-analyse the findings from eligible studies.

RESULTS

We included 17 unique studies with a total of 47 374 child participants. Maternal and paternal substance use were both associated with increased odds of child any drug use [OR = 2.09; 95% confidence interval (CI) = 1.53, 2.86; n = 12 349 participants; three studies and OR = 2.86; 95% CI = 1.25, 6.54; n = 5692 participants; three studies, respectively], child alcohol problem use (OR = 2.16; 95% CI = 1.73, 2.71; n = 7339 participants; four studies and OR = 1.70; 95% CI = 1.36, 2.12; n = 14 219 participants; six studies), child externalizing problems (OR = 1.81; 95% CI = 1.01, 3.22; n = 1748 participants; three studies and OR = 1.60; 95% CI = 1.18, 2.17; n = 2508 participants; six studies) and child internalizing problems (OR = 1.60; 95% CI = 1.25, 2.06; n = 1748 participants; three studies and OR = 1.42; 95% CI = 1.12, 1.81; n = 2248 participants; five studies). Child any alcohol use was associated with maternal substance use only (OR = 2.26; 95% CI = 1.08, 4.70; n = 28 691 participants; five studies).

CONCLUSIONS

Both maternal and paternal substance use are associated with child substance use and mental health problems.

Can we mitigate the psychological impacts of social isolation using behavioural activation? Long-term results of the UK BASIL urgent public health COVID-19 pilot randomised controlled trial and living systematic review

BACKGROUND

Behavioural and cognitive interventions remain credible approaches in addressing loneliness and depression. There was a need to rapidly generate and assimilate trial-based data during COVID-19.

OBJECTIVES

We undertook a parallel pilot RCT of behavioural activation (a brief behavioural intervention) for depression and loneliness (Behavioural Activation in Social Isolation, the BASIL-C19 trial ISRCTN94091479). We also assimilate these data in a living systematic review (PROSPERO CRD42021298788) of cognitive and/or behavioural interventions.

METHODS

Participants (≥65 years) with long-term conditions were computer randomised to behavioural activation (n=47) versus care as usual (n=49). Primary outcome was PHQ-9. Secondary outcomes included loneliness (De Jong Scale). Data from the BASIL-C19 trial were included in a metanalysis of depression and loneliness.

FINDINGS

The 12 months adjusted mean difference for PHQ-9 was -0.70 (95% CI -2.61 to 1.20) and for loneliness was -0.39 (95% CI -1.43 to 0.65).The BASIL-C19 living systematic review (12 trials) found short-term reductions in depression (standardised mean difference (SMD)=-0.31, 95% CI -0.51 to -0.11) and loneliness (SMD=-0.48, 95% CI -0.70 to -0.27). There were few long-term trials, but there was evidence of some benefit (loneliness SMD=-0.20, 95% CI -0.40 to -0.01; depression SMD=-0.20, 95% CI -0.47 to 0.07).

DISCUSSION

We delivered a pilot trial of a behavioural intervention targeting loneliness and depression; achieving long-term follow-up. Living meta-analysis provides strong evidence of short-term benefit for loneliness and depression for cognitive and/or behavioural approaches. A fully powered BASIL trial is underway.

CLINICAL IMPLICATIONS

Scalable behavioural and cognitive approaches should be considered as population-level strategies for depression and loneliness on the basis of a living systematic review.

The development and validation of a prognostic model to PREDICT Relapse of depression in adult patients in primary care: protocol for the PREDICTR study

BACKGROUND

Most patients who present with depression are treated in primary care by general practitioners (GPs). Relapse of depression is common (at least 50% of patients treated for depression will relapse after a single episode) and leads to considerable morbidity and decreased quality of life for patients. The majority of patients will relapse within 6 months, and those with a history of relapse are more likely to relapse in the future than those with no such history. GPs see a largely undifferentiated case-mix of patients, and once patients with depression reach remission, there is limited guidance to help GPs stratify patients according to risk of relapse. We aim to develop a prognostic model to predict an individual's risk of relapse within 6-8 months of entering remission. The long-term objective is to inform the clinical management of depression after the acute phase.

METHODS

We will develop a prognostic model using secondary analysis of individual participant data drawn from seven RCTs and one longitudinal cohort study in primary or community care settings. We will use logistic regression to predict the outcome of relapse of depression within 6-8 months. We plan to include the following established relapse predictors in the model: residual depressive symptoms, number of previous depressive episodes, co-morbid anxiety and severity of index episode. We will use a "full model" development approach, including all available predictors. Performance statistics (optimism-adjusted C-statistic, calibration-in-the-large, calibration slope) and calibration plots (with smoothed calibration curves) will be calculated. Generalisability of predictive performance will be assessed through internal-external cross-validation. Clinical utility will be explored through net benefit analysis.

DISCUSSION

We will derive a statistical model to predict relapse of depression in remitted depressed patients in primary care. Assuming the model has sufficient predictive performance, we outline the next steps including independent external validation and further assessment of clinical utility and impact.

STUDY REGISTRATION

ClinicalTrials.gov ID: NCT04666662.

Interventions for adults with a history of complex traumatic events: the INCiTE mixed-methods systematic review

BACKGROUND

People with a history of complex traumatic events typically experience trauma and stressor disorders and additional mental comorbidities. It is not known if existing evidence-based treatments are effective and acceptable for this group of people.

OBJECTIVE

To identify candidate psychological and non-pharmacological treatments for future research.

DESIGN

Mixed-methods systematic review.

PARTICIPANTS

Adults aged ≥ 18 years with a history of complex traumatic events.

INTERVENTIONS

Psychological interventions versus control or active control; pharmacological interventions versus placebo.

MAIN OUTCOME MEASURES

Post-traumatic stress disorder symptoms, common mental health problems and attrition.

DATA SOURCES

Cumulative Index to Nursing and Allied Health Literature (CINAHL) (1937 onwards); Cochrane Central Register of Controlled Trials (CENTRAL) (from inception); EMBASE (1974 to 2017 week 16); International Pharmaceutical Abstracts (1970 onwards); MEDLINE and MEDLINE Epub Ahead of Print and In-Process & Other Non-Indexed Citations (1946 to present); Published International Literature on Traumatic Stress (PILOTS) (1987 onwards); PsycINFO (1806 to April week 2 2017); and Science Citation Index (1900 onwards). Searches were conducted between April and August 2017.

REVIEW METHODS

Eligible studies were singly screened and disagreements were resolved at consensus meetings. The risk of bias was assessed using the Cochrane risk-of-bias tool and a bespoke version of a quality appraisal checklist used by the National Institute for Health and Care Excellence. A meta-analysis was conducted across all populations for each intervention category and for population subgroups. Moderators of effectiveness were assessed using metaregression and a component network meta-analysis. A qualitative synthesis was undertaken to summarise the acceptability of interventions with the relevance of findings assessed by the GRADE-CERQual checklist.

RESULTS

One hundred and four randomised controlled trials and nine non-randomised controlled trials were included. For the qualitative acceptability review, 4324 records were identified and nine studies were included. The population subgroups were veterans, childhood sexual abuse victims, war affected, refugees and domestic violence victims. Psychological interventions were superior to the control post treatment for reducing post-traumatic stress disorder symptoms (standardised mean difference -0.90, 95% confidence interval -1.14 to -0.66; number of trials = 39) and also for associated symptoms of depression, but not anxiety. Trauma-focused therapies were the most effective interventions across all populations for post-traumatic stress disorder and depression. Multicomponent and trauma-focused interventions were effective for negative self-concept. Phase-based approaches were also superior to the control for post-traumatic stress disorder and depression and showed the most benefit for managing emotional dysregulation and interpersonal problems. Only antipsychotic medication was effective for reducing post-traumatic stress disorder symptoms; medications were not effective for mental comorbidities. Eight qualitative studies were included. Interventions were more acceptable if service users could identify benefits and if they were delivered in ways that accommodated their personal and social needs.

LIMITATIONS

Assessments about long-term effectiveness of interventions were not possible. Studies that included outcomes related to comorbid psychiatric states, such as borderline personality disorder, and populations from prisons and humanitarian crises were under-represented.

CONCLUSIONS

Evidence-based psychological interventions are effective and acceptable post treatment for reducing post-traumatic stress disorder symptoms and depression and anxiety in people with complex trauma. These interventions were less effective in veterans and had less of an impact on symptoms associated with complex post-traumatic stress disorder.

FUTURE WORK

Definitive trials of phase-based versus non-phase-based interventions with long-term follow-up for post-traumatic stress disorder and associated mental comorbidities.

STUDY REGISTRATION

This study is registered as PROSPERO CRD42017055523.

FUNDING

This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 24, No. 43. See the NIHR Journals Library website for further project information.

A systematic review and meta-analysis of the prevalence of common mental disorders in people with non-communicable diseases in Bangladesh, India, and Pakistan

Background

The prevalence of mental and physical comorbidities is unknown in South Asia, as estimates of mental ill health in patients with non-communicable diseases (NCDs) have predominantly come from studies based in the United States, Europe and Australasia. This systematic review and meta-analysis summarises evidence and provides pooled estimates of the prevalence of common mental disorders in adults with non-communicable diseases in South Asia.

Methods

We included prevalence studies of depression and anxiety in adults with diabetes, cancer, cardiovascular disease, and chronic respiratory conditions in Bangladesh, India, and Pakistan, published from 1990 onwards in international and country-specific databases.

Results

Out of 96 included studies, 83 provided data for random effects meta-analyses. The pooled prevalence of depression was 44% (95% confidence interval (CI) = 26 to 62) for patients with COPD, 40% (95% CI = 34 to 45) for diabetes, 39% (95% CI = 23 to 56) for stroke, 38% (95% CI = 32 to 45) for hypertension, and 37% (95% CI = 30 to 45) for cancer. The pooled prevalence of anxiety based on 28 studies was 29% (95% CI = 22 to 36). Many quality issues were identified in a critical appraisal of included studies, mostly relating to the sampling frame and selection process, the description of the methods and basic data, and the description of non-responders.

Conclusions

Depression and anxiety are prevalent and underdiagnosed in people with physical comorbidities in Bangladesh, India, and Pakistan.

Assessment of C-Reactive Protein Diagnostic Test Accuracy for Late-Onset Infection in Newborn Infants: A Systematic Review and Meta-analysis

IMPORTANCE

Rapid and accurate diagnosis of late-onset infection in newborn infants could inform treatment decisions and avoid unnecessary administration of antibiotics.

OBJECTIVE

To compare the accuracy of serum C-reactive protein (CRP) with that of microbiological blood culture for diagnosing late-onset infection in newborns.

DATA SOURCES

MEDLINE (1946-2019), Embase (1946-2019), and Science Citation Index (1900-2019) databases were searched for references (any language). The MeSH search terms included were "exp infant, newborn/" or "premature birth/" plus free text synonyms; and "C-reactive protein/" plus free text synonyms; and "exp sepsis/" or "exp bacterial infections/" plus free text synonyms. The proceedings from relevant conferences and references of identified papers were scrutinized. Authors were contacted to request missing data.

STUDY SELECTION

Cohort and cross-sectional studies were included that compared the accuracy of serum CRP levels with microbiological culture results to diagnose late-onset (>72 hours after birth) infection in newborns (any gestational age) hospitalized after birth. Two reviewers assessed study eligibility. Among 10 394 records, 148 studies were assessed as full texts.

DATA EXTRACTION AND SYNTHESIS

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guideline extension for Diagnostic Test Accuracy (DTA) reviews was followed. Two reviewers assessed the method quality of each study using guidance from the Cochrane Screening and Diagnostic Test Methods Group (adapted from the Quality Assessment of Diagnostic Accuracy Studies 2).

MAIN OUTCOMES AND MEASURES

The primary meta-analysis outcome was diagnostic test accuracy of serum CRP level taken at initial investigation of an infant with suspected late-onset infection. The median specificity (proportion of true-negative results) and calculated pooled sensitivity (proportion of true-positive results) were determined by generating hierarchical summary receiver characteristic operating curves.

RESULTS

In total, 22 studies with 2255 infants were included (sample size range, 11-590 infants). Participants in most studies were preterm (<37 weeks) or very low-birth weight (<1500 g) infants. Two studies additionally enrolled infants born at term. Most studies (16) used a prespecified CRP level cutoff for a "positive" index test (5-10 mg/L), and most studies (17) used the culture of a pathogenic microorganism from blood as the reference standard. Risk of bias was low with independent assessment of index and reference tests. At median specificity (0.74), pooled sensitivity was 0.62 (95% CI, 0.50-0.72). Adding serum CRP level to the assessment of an infant with a 40% pretest probability of late-onset infection (the median for the included studies) generated posttest probabilities of 26% for a negative test result and 61% for a positive test result.

CONCLUSIONS AND RELEVANCE

The findings suggest that determination of serum CRP level at initial evaluation of an infant with suspected late-onset infection is unlikely to aid early diagnosis or to select infants to undergo further investigation or treatment with antimicrobial therapy or other interventions.

The public health evidence-base on novel psychoactive substance use: scoping review with narrative synthesis of selected bodies of evidence

Background

This review aimed to address what was known about the public health burden associated with novel psychoactive substances (NPS) use, and the effectiveness of responses targeting NPS use and/or associated problems.

Methods

Relevant literature was identified through a range of searches covering the period from January 2006 to June 2016. Data synthesis was undertaken in three stages. Firstly we mapped the evidence available in order to characterize the literature according to a set of analytic categories developed a priori. Secondly, we identified evidence gaps from a set of a priori research questions. Finally, we then undertook a narrative synthesis of selected bodies of evidence, interpreting data using a conceptual framework specifically designed for use in this review.

Results

A total of 995 articles were included in the scoping review with the majority being case reports/series on individual level adverse effects due to NPS use. We synthesized UK data from 29 surveys and 7 qualitative studies, and international data in 10 systematic reviews on harms associated with NPS use, and 17 evaluations of policy responses. We found little data on risk factors, harms associated with long-term NPS use, and interventions.

Conclusion

In all cases we found the available evidence to be at an early stage of development.

Strimvelis® for Treating Severe Combined Immunodeficiency Caused by Adenosine Deaminase Deficiency: An Evidence Review Group Perspective of a NICE Highly Specialised Technology Evaluation

The Centre for Reviews and Dissemination and Centre for Health Economics Technology Assessment Group at the University of York was commissioned by the National Institute for Health and Care Excellence (NICE) Highly Specialised Technologies (HST) programme to act as the independent Evidence Review Group (ERG) for an appraisal of Strimvelis^®, a gene therapy treatment for adenosine deaminase deficiency-severe combined immunodeficiency (ADA-SCID). This paper describes the manufacturing company's submission of clinical and economic evidence, the ERG's review and the resulting NICE guidance. For Strimvelis^® compared with haematopoietic stem cell transplant (HSCT) from a matched unrelated donor (MUD) and HSCT from a haploidentical donor, the company base-case deterministic incremental cost-effectiveness ratios (ICERs) were £36,360 and £14,645 per quality-adjusted life-year (QALY) gained, respectively (using a discount rate of 1.5%). Although overall survival in patients receiving Strimvelis^® was substantially higher than historical comparator data on HSCT from a MUD or haploidentical donor, the ERG was concerned that the estimated treatment benefit remained highly uncertain. The ERG critiqued some assumptions in the cost-effectiveness model, including that all patients return to general population mortality and morbidity after a successful procedure; that all patients receive a matched sibling donor following an unsuccessful engraftment; and that differences in wait times exist between the treatments. Incorporating a number of changes to the model, the ERG's base-case ICERs were £86,815 per QALY gained for Strimvelis^® compared with HSCT from a MUD and £16,704 per QALY gained compared with HSCT from a haploidentical donor (using a discount rate of 1.5%). The ICER for Strimvelis^® compared with HSCT from a MUD was highly sensitive to the difference in procedural mortality and could exceed NICE's £100,000 per QALY gained threshold for HSTs, if HSCT survival rates have improved since the most recent data. The evaluation committee concluded that the most plausible ICERs were lower than £100,000 per QALY gained and that Strimvelis^® should be recommended for treatment of ADA-SCID where a matched related donor is unavailable.

Sexual orientation and suicidal behaviour in young people

Lesbian, gay, and bisexual (LGB) young people have been found to be at greater risk of suicidal behaviour. National prevention strategies have identified the need to reduce suicide risk in this population. However, research on specific risk factors for LGB young people that might inform suicide prevention programmes are at an early stage of development.

Multiple Risk Behavior Interventions: Meta-analyses of RCTs

CONTEXT

Multiple risk behaviors are common and associated with developing chronic conditions such as heart disease, cancer, or Type 2 diabetes. A systematic review, meta-analysis, and meta-regression of the effectiveness of multiple risk behavior interventions was conducted.

EVIDENCE ACQUISITION

Six electronic databases including MEDLINE, EMBASE, and PsycINFO were searched to August 2016. RCTs of non-pharmacologic interventions in general adult populations were selected. Studies targeting specific at-risk groups (such as people screened for cardiovascular risk factors or obesity) were excluded. Studies were screened independently. Study characteristics and outcomes were extracted and risk of bias assessed by one researcher and checked by another. The Behaviour Change Wheel and Oxford Implementation Index were used to code intervention content and context.

EVIDENCE SYNTHESIS

Random-effects meta-analyses were conducted. Sixty-nine trials involving 73,873 individuals were included. Interventions mainly comprised education and skills training and were associated with modest improvements in most risk behaviors: increased fruit and vegetable intake (0.31 portions, 95% CI=0.17, 0.45) and physical activity (standardized mean difference, 0.25; 95% CI=0.13, 0.38), and reduced fat intake (standardized mean difference, -0.24; 95% CI=-0.36, -0.12). Although reductions in smoking were found (OR=0.78, 95% CI=0.68, 0.90), they appeared to be negatively associated with improvement in other behaviors (such as diet and physical activity). Preliminary evidence suggests that sequentially changing smoking alongside other risk behaviors was more effective than simultaneous change. But most studies assessed simultaneous rather than sequential change in risk behaviors; therefore, comparisons are sparse. Follow-up period and intervention characteristics impacted effectiveness for some outcomes.

CONCLUSIONS

Interventions comprising education (e.g., providing information about behaviors associated with health risks) and skills training (e.g., teaching skills that equip participants to engage in less risky behavior) and targeting multiple risk behaviors concurrently are associated with small changes in diet and physical activity. Although on average smoking was reduced, it appeared changes in smoking were negatively associated with changes in other behaviors, suggesting it may not be optimal to target smoking simultaneously with other risk behaviors.

Prevalence and predictors of post-stroke mood disorders: A meta-analysis and meta-regression of depression, anxiety and adjustment disorder

OBJECTIVE

To ascertain the prevalence and predictors of mood disorders, determined by structured clinical interviews (ICD or DSM criteria) in people after stroke.

METHODS

Major electronic databases were searched from inception to June 2016 for studies involving major depression (MDD), minor depression (MnD), dysthymia, adjustment disorder, any depressive disorder (any depressive disorder) and anxiety disorders. Studies were combined using both random and fixed effects meta-analysis and results were stratified as appropriate.

RESULTS

Depression was examined on 147 occasions from 2days to 7years after stroke (mean 6.87months, N=33 in acute, N=43 in rehabilitation and N=69 in the community/outpatients). Across 128 analyses involving 15,573 patients assessed for major depressive disorder (MDD), the point prevalence of depression was 17.7% (95% CI=15.6% to 20.0%) 0.65 analyses involving 9720 patients determined MnD was present in 13.1% in all settings (95% CI=10.9% to 15.8%). Dysthymia was present in 3.1% (95% CI=2.1% to 5.3%), adjustment disorder in 6.9% (95% CI=4.6 to 9.7%) and anxiety in 9.8% (95% CI=5.9% to 14.8%). Any depressive disorder was present in 33.5% (95% CI=30.3% to 36.8%). The relative risk of any depressive disorder was higher following left (dominant) hemisphere stroke, aphasia, and among people with a family history and past history of mood disorders.

CONCLUSION

Depression, adjustment disorder and anxiety are common after stroke. Risk factors are aphasia, dominant hemispheric lesions and past personal/family history of depression but not time since stroke.

The effect of Liver Transplantation on the quality of life of the recipient's main caregiver - a systematic review

Liver transplantation (LT) is a transformative, life-saving procedure with life-long sequale for patients and their caregivers. The impact of LT on the patient's main caregiver can be underestimated. We carried out a systematic review of the impact of LT on the Health-Related Quality of Life (HRQL) of LT patients' main caregivers. We searched 13 medical databases from 1996 to 2015. We included studies with HRQL data on caregivers of patients following LT then quality assessed and narratively synthesized the findings from these studies. Of 7076 initial hits, only five studies fell within the scope of this study. In general, they showed caregiver burden persisted in the early period following LT. One study showed improvements, however, the other four showed caregiver's levels of stress, anxiety and depression, remained similar or got worse post-LT and remained above that of the normal population. It was suggested that HRQL of the patient impacted on the caregiver and vice versa and may be linked to patient outcomes. No data were available investigating which groups were at particular risk of low HRQL following LT or if any interventions could improve this. The current information about LT caregivers' needs and factors that impact on their HRQL are not adequately defined. Large studies are needed to examine the effects of LT on the patients' family and caregivers to understand the importance of caregiver support to maximize outcomes of LT for the patient and their caregivers.

The Novel Psychoactive Substances in the UK Project: empirical and conceptual review work to produce research recommendations

Background

Although illegal drug use has largely been declining in the UK over the past decade, this period has witnessed the emergence of a range of novel psychoactive substances (NPS) (‘legal highs’). These are new, mostly synthetic, substances that mimic the effects of existing drugs). Despite there being many causes for concern in relation to NPS, there has been little prior study of the burden associated with their use in public health terms. Clarity is lacking on research priorities in this rapidly developing literature.

Objectives

To inform the development of public health intervention research on NPS by reviewing existing data on their use, associated problems and potential responses to such problems.

Design

A scoping review and narrative synthesis of selected bodies of evidence was undertaken to summarise and evaluate what is known about NPS use and the related harms of, and responses to, such use. Relevant literature was identified from electronic databases (covering January 2006 to June 2016 inclusive), Google (Google Inc., Mountain View, CA, USA), relevant websites and online drug forums and by contacting experts. Articles were included if they were primary studies, secondary studies involving the analysis and interpretation of primary research or discussion papers. A conceptual framework postulating an evidence-informed public health approach to NPS use in the UK was developed through a pragmatic literature review, the iterative development of concepts and finalisation in light of the results from the empirical review work. The process also involved feedback from various stakeholders. Research recommendations were developed from both strands of work.

Results

A total of 995 articles were included in the scoping review, the majority of which related to individual-level health-related adverse effects attributable to NPS use. The prevalence of lifetime NPS use varied widely between (e.g. with higher prevalence in young males) and within population subgroups. The most commonly reported adverse effects were psychiatric/other neurological, cardiovascular, renal and gastrointestinal manifestations, and there is limited evidence available on responses. In these and other respects, available evidence is at an early stage of development. Initial evidence challenges the view that NPS should be treated differently from other illicit drugs. The conceptual framework indicated that much of the evidence that would be useful to inform public health responses does not yet exist. We propose a systems-based prevention approach that develops existing responses, is multilevel and life course informed in character, and emphasises commonalities between NPS and other legal and illegal drug use. We make 20 recommendations for research, including nine key recommendations.

Limitations

Scoping reviews do not interrogate evidence in depth, and the disjunction between the scoping review and the conceptual framework findings is worthy of careful attention.

Conclusions

Key research recommendations build on those that have previously been made and offer more evidence-based justification and detail, as previous recommendations have not yet been acted on. The case for decision-making on commissioning new research based on these recommendations is both strong and urgent.

Future work

The validity of recommendations generated through this project could be enhanced via further work with research commissioners, policy-makers, researchers and the public.

Study registration

The systematic review element of this study is registered as PROSPERO CRD42016026415.

Funding

The National Institute for Health Research Public Health Research programme.

Promising evidence for nutrition interventions in people with serious mental illness

People with serious mental illness (SMI) are more likely to engage in health risk behaviours such as unhealthy eating, physical inactivity and smoking. The review by Teasdale et al in this issue shows the potential for nutrition interventions to help people with SMI to manage their weight.

Improving outcomes for people in mental health crisis: a rapid synthesis of the evidence for available models of care

BACKGROUND

Crisis Concordat was established to improve outcomes for people experiencing a mental health crisis. The Crisis Concordat sets out four stages of the crisis care pathway: (1) access to support before crisis point; (2) urgent and emergency access to crisis care; (3) quality treatment and care in crisis; and (4) promoting recovery.

OBJECTIVES

To evaluate the clinical effectiveness and cost-effectiveness of the models of care for improving outcomes at each stage of the care pathway.

DATA SOURCES

Electronic databases were searched for guidelines, reviews and, where necessary, primary studies. The searches were performed on 25 and 26 June 2014 for NHS Evidence, Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, NHS Economic Evaluation Database, and the Health Technology Assessment (HTA) and PROSPERO databases, and on 11 November 2014 for MEDLINE, PsycINFO and the Criminal Justice Abstracts databases. Relevant reports and reference lists of retrieved articles were scanned to identify additional studies.

STUDY SELECTION

When guidelines covered a topic comprehensively, further literature was not assessed; however, where there were gaps, systematic reviews and then primary studies were assessed in order of priority.

STUDY APPRAISAL AND SYNTHESIS METHODS

Systematic reviews were critically appraised using the Risk Of Bias In Systematic reviews assessment tool, trials were assessed using the Cochrane risk-of-bias tool, studies without a control group were assessed using the National Institute for Health and Care Excellence (NICE) prognostic studies tool and qualitative studies were assessed using the Critical Appraisal Skills Programme quality assessment tool. A narrative synthesis was conducted for each stage of the care pathway structured according to the type of care model assessed. The type and range of evidence identified precluded the use of meta-analysis.

RESULTS AND LIMITATIONS

One review of reviews, six systematic reviews, nine guidelines and 15 primary studies were included. There was very limited evidence for access to support before crisis point. There was evidence of benefits for liaison psychiatry teams in improving service-related outcomes in emergency departments, but this was often limited by potential confounding in most studies. There was limited evidence regarding models to improve urgent and emergency access to crisis care to guide police officers in their Mental Health Act responsibilities. There was positive evidence on clinical effectiveness and cost-effectiveness of crisis resolution teams but variability in implementation. Current work from the Crisis resolution team Optimisation and RElapse prevention study aims to improve fidelity in delivering these models. Crisis houses and acute day hospital care are also currently recommended by NICE. There was a large evidence base on promoting recovery with a range of interventions recommended by NICE likely to be important in helping people stay well.

CONCLUSIONS AND IMPLICATIONS

Most evidence was rated as low or very low quality, but this partly reflects the difficulty of conducting research into complex interventions for people in a mental health crisis and does not imply that all research was poorly conducted. However, there are currently important gaps in research for a number of stages of the crisis care pathway. Particular gaps in research on access to support before crisis point and urgent and emergency access to crisis care were found. In addition, more high-quality research is needed on the clinical effectiveness and cost-effectiveness of mental health crisis care, including effective components of inpatient care, post-discharge transitional care and Community Mental Health Teams/intensive case management teams.

STUDY REGISTRATION

This study is registered as PROSPERO CRD42014013279.

FUNDING

The National Institute for Health Research HTA programme.

Predicting suicide following self-harm: systematic review of risk factors and risk scales

BACKGROUND

People with a history of self-harm are at a far greater risk of suicide than the general population. However, the relationship between self-harm and suicide is complex.

AIMS

To undertake the first systematic review and meta-analysis of prospective studies of risk factors and risk assessment scales to predict suicide following self-harm.

METHOD

We conducted a search for prospective cohort studies of populations who had self-harmed. For the review of risk scales we also included studies examining the risk of suicide in people under specialist mental healthcare, in order to broaden the scope of the review and increase the number of studies considered. Differences in predictive accuracy between populations were examined where applicable.

RESULTS

Twelve studies on risk factors and 7 studies on risk scales were included. Four risk factors emerged from the meta-analysis, with robust effect sizes that showed little change when adjusted for important potential confounders. These included: previous episodes of self-harm (hazard ratio (HR) = 1.68, 95% CI 1.38-2.05, K = 4), suicidal intent (HR = 2.7, 95% CI 1.91-3.81, K = 3), physical health problems (HR = 1.99, 95% CI 1.16-3.43, K = 3) and male gender (HR = 2.05, 95% CI 1.70-2.46, K = 5). The included studies evaluated only three risk scales (Beck Hopelessness Scale (BHS), Suicide Intent Scale (SIS) and Scale for Suicide Ideation). Where meta-analyses were possible (BHS, SIS), the analysis was based on sparse data and a high heterogeneity was observed. The positive predictive values ranged from 1.3 to 16.7%.

CONCLUSIONS

The four risk factors that emerged, although of interest, are unlikely to be of much practical use because they are comparatively common in clinical populations. No scales have sufficient evidence to support their use. The use of these scales, or an over-reliance on the identification of risk factors in clinical practice, may provide false reassurance and is, therefore, potentially dangerous. Comprehensive psychosocial assessments of the risks and needs that are specific to the individual should be central to the management of people who have self-harmed.

A systematic review on the clustering and co-occurrence of multiple risk behaviours

BACKGROUND

Risk behaviours, such as smoking and physical inactivity account for up to two-thirds of all cardiovascular deaths, and are associated with substantial increased mortality in many conditions including cancer and diabetes. As risk behaviours are thought to co-occur in individuals we conducted a systematic review of studies addressing clustering or co-occurrence of risk behaviours and their predictors. As the main aim of the review was to inform public health policy in England we limited inclusion to studies conducted in the UK.

METHODS

Key databases were searched from 1990 to 2016. We included UK based cross-sectional and longitudinal studies that investigated risk behaviours such as smoking, physical inactivity, unhealthy diet. High heterogeneity precluded meta-analyses.

RESULTS

Thirty-seven studies were included in the review (32 cross-sectional and five longitudinal). Most studies investigated unhealthy diet, physical inactivity, alcohol misuse, and smoking. In general adult populations, there was relatively strong evidence of clustering between alcohol misuse and smoking; and unhealthy diet and smoking. For young adults, there was evidence of clustering between sexual risk behaviour and smoking, sexual risk behaviour and illicit drug use, and sexual risk behaviour and alcohol misuse. The strongest associations with co-occurrence and clustering of multiple risk behaviours were occupation (up to 4-fold increased odds in lower SES groups) and education (up to 5-fold increased odds in those with no qualifications).

CONCLUSIONS

Among general adult populations, alcohol misuse and smoking was the most commonly identified risk behaviour cluster. Among young adults, there was consistent evidence of clustering found between sexual risk behaviour and substance misuse. Socio-economic status was the strongest predictor of engaging in multiple risk behaviours. This suggests the potential for interventions targeting multiple risk behaviours either sequentially or concurrently particularly where there is evidence of clustering. In addition, there is potential for intervening at the social or environmental level due to the strong association with socio-economic status.

The Use of Bayesian Networks to Assess the Quality of Evidence from Research Synthesis: 2. Inter-Rater Reliability and Comparison with Standard GRADE Assessment

Background

The grades of recommendation, assessment, development and evaluation (GRADE) approach is widely implemented in systematic reviews, health technology assessment and guideline development organisations throughout the world. We have previously reported on the development of the Semi-Automated Quality Assessment Tool (SAQAT), which enables a semi-automated validity assessment based on GRADE criteria. The main advantage to our approach is the potential to improve inter-rater agreement of GRADE assessments particularly when used by less experienced researchers, because such judgements can be complex and challenging to apply without training. This is the first study examining the inter-rater agreement of the SAQAT.

Methods

We conducted two studies to compare: a) the inter-rater agreement of two researchers using the SAQAT independently on 28 meta-analyses and b) the inter-rater agreement between a researcher using the SAQAT (who had no experience of using GRADE) and an experienced member of the GRADE working group conducting a standard GRADE assessment on 15 meta-analyses.

Results

There was substantial agreement between independent researchers using the Quality Assessment Tool for all domains (for example, overall GRADE rating: weighted kappa 0.79; 95% CI 0.65 to 0.93). Comparison between the SAQAT and a standard GRADE assessment suggested that inconsistency was parameterised too conservatively by the SAQAT. Therefore the tool was amended. Following amendment we found fair-to-moderate agreement between the standard GRADE assessment and the SAQAT (for example, overall GRADE rating: weighted kappa 0.35; 95% CI 0.09 to 0.87).

Conclusions

Despite a need for further research, the SAQAT may aid consistent application of GRADE, particularly by less experienced researchers.

The use of Bayesian networks to assess the quality of evidence from research synthesis: 1

Background

The grades of recommendation, assessment, development and evaluation (GRADE) approach is widely implemented in systematic reviews, health technology assessment and guideline development organisations throughout the world. A key advantage to this approach is that it aids transparency regarding judgments on the quality of evidence. However, the intricacies of making judgments about research methodology and evidence make the GRADE system complex and challenging to apply without training.

Methods

We have developed a semi-automated quality assessment tool (SAQAT) l based on GRADE. This is informed by responses by reviewers to checklist questions regarding characteristics that may lead to unreliability. These responses are then entered into the Bayesian network to ascertain the probabilities of risk of bias, inconsistency, indirectness, imprecision and publication bias conditional on review characteristics. The model then combines these probabilities to provide a probability for each of the GRADE overall quality categories. We tested the model using a range of plausible scenarios that guideline developers or review authors could encounter.

Results

Overall, the model reproduced GRADE judgements for a range of scenarios. Potential advantages over standard assessment are use of explicit and consistent weightings for different review characteristics, forcing consideration of important but sometimes neglected characteristics and principled downgrading where small but important probabilities of downgrading are accrued across domains.

Conclusions

Bayesian networks have considerable potential for use as tools to assess the validity of research evidence. The key strength of such networks lies in the provision of a statistically coherent method for combining probabilities across a complex framework based on both belief and evidence. In addition to providing tools for less experienced users to implement reliability assessment, the potential for sensitivity analyses and automation may be beneficial for application and the methodological development of reliability tools.

Characteristics of interventions targeting multiple lifestyle risk behaviours in adult populations: a systematic scoping review

BACKGROUND

Modifiable lifestyle risk behaviours such as smoking, unhealthy diet, physical inactivity and alcohol misuse are the leading causes of major, non-communicable diseases worldwide. It is increasingly being recognised that interventions which target more than one risk behaviour may be an effective and efficient way of improving people's lifestyles. To date, there has been no attempt to summarise the global evidence base for interventions targeting multiple risk behaviours.

OBJECTIVE

To identify and map the characteristics of studies evaluating multiple risk behaviour change interventions targeted at adult populations in any country.

METHODS

Seven bibliographic databases were searched between January, 1990, and January/ May, 2013. Authors of protocols, conference abstracts, and other relevant articles were contacted. Study characteristics were extracted and inputted into Eppi-Reviewer 4.

RESULTS

In total, 220 studies were included in the scoping review. Most were randomised controlled trials (62%) conducted in the United States (49%), and targeted diet and physical activity (56%) in people from general populations (14%) or subgroups of general populations (45%). Very few studies had been conducted in the Middle East (2%), Africa (0.5%), or South America (0.5%). There was also a scarcity of studies conducted among young adults (1%), or racial and minority ethnic populations (4%) worldwide.

CONCLUSIONS

Research is required to investigate the interrelationships of lifestyle risk behaviours in varying cultural contexts around the world. Cross-cultural development and evaluation of multiple risk behaviour change interventions is also needed, particularly in populations of young adults and racial and minority ethnic populations.

A checklist designed to aid consistency and reproducibility of GRADE assessments: development and pilot validation

BACKGROUND

The grading of recommendation, assessment, development and evaluation (GRADE) approach is widely implemented in health technology assessment and guideline development organisations throughout the world. GRADE provides a transparent approach to reaching judgements about the quality of evidence on the effects of a health care intervention, but is complex and therefore challenging to apply in a consistent manner.

METHODS

We developed a checklist to guide the researcher to extract the data required to make a GRADE assessment. We applied the checklist to 29 meta-analyses of randomised controlled trials on the effectiveness of health care interventions. Two reviewers used the checklist for each paper and used these data to rate the quality of evidence for a particular outcome.

RESULTS

For most (70%) checklist items, there was good agreement between reviewers. The main problems were for items relating to indirectness where considerable judgement is required.

CONCLUSIONS

There was consistent agreement between reviewers on most items in the checklist. The use of this checklist may be an aid to improving the consistency and reproducibility of GRADE assessments, particularly for inexperienced users or in rapid reviews without the resources to conduct assessments by two researchers independently.

Accuracy of one or two simple questions to identify alcohol-use disorder in primary care: a meta-analysis

BACKGROUND

There is much interest in ultra-short alcohol screening in primary care that may support brief alcohol interventions. Brief screening consisting of one or two questions might be used alone or in combination with longer tests as recommended by the Primary Care Service Framework.

AIM

To investigate whether a simple one and two question screening might prove an accurate and acceptable screening method in primary care.

DESIGN AND SETTING

A systematic literature search, critical appraisal and meta-analysis were conducted.

METHOD

A comprehensive search identified 61 analyses of single questions to detect alcohol problems including 17 that took place in primary care, using a robust interview standard. Despite focusing alcohol-use disorder in primary care settings, heterogeneity remained high, therefore random effects and bivariate meta-analyses were used.

RESULTS

After adjustments, diagnostic accuracy of a single-question approach was given by a sensitivity of 54.5% (95% CI = 43.0% to 65.5%) and a specificity of 87.3% (95% CI = 81.5% to 91.5%) using meta-analytic weighting. Two questions had a sensitivity of 87.2% (95% CI = 69.9% to 97.7%) and specificity of 79.8% (95% CI = 75.7% to 83.6%). Looking at each question individually, the most successful single question was a modification of the Single Alcohol Screening Question (SASQ) namely, 'How often do you have six or more drinks on one occasion?'. The optimal approach appears to be two questions followed by the CAGE questionnaire, which achieved an overall accuracy of 90.9% and required only 3.3 questions per attendee.

CONCLUSION

Two brief questions can be used as an initial screen for alcohol problems but only when combined with a second-step screen. A brief alcohol intervention should be considered in those individuals who answer positively on both steps.

Screening for poststroke major depression: a meta-analysis of diagnostic validity studies

BACKGROUND

Major depression is common in stroke patients and associated with increased rates of disability and mortality. Identifying depression may improve mental and physical health. The aim of this review was to determine the most accurate tool for detecting poststroke depression.

METHODS

Seven databases were searched up to November 2012. Two authors selected studies using International Classification of Disease  or Diagnostic and Statistical Manual diagnosis of depression as the reference standard. Two authors extracted data and assessed methodological quality. Included studies were synthesised using meta-analyses.

RESULTS

A total of 24 included studies provided data on 2907 participants. The Center of Epidemiological Studies-Depression Scale (CESD) (sensitivity: 0.75; 95% CI 0.60 to 0.85; specificity: 0.88; 95% CI 0.71 to 0.95), the Hamilton Depression Rating Scale (HDRS) (sensitivity: 0.84; 95% CI 0.75 to 0.90; specificity:0.83; 95% CI 0.72 to 0.90) and the Patient Health Questionnaire (PHQ)-9 (sensitivity: 0.86; 95% CI 0.70 to 0.94; specificity: 0.79; 95% CI 0.60 to 0.90) appeared to be the optimal measures for screening measures. However, the clinical utility of all tools was modest for case-finding.

INTERPRETATION

There are a number of possible instruments that may help in screening for poststroke depression but none are satisfactory for case-finding. Preliminary data suggests the CESD, HDRS or the PHQ-9 as the most promising options. Although it should be noted such scales should not be used in isolation but followed up with a more detailed clinical assessment. While there is promising data for the PHQ-2 in other populations, it performed less well than other measures.

Potential uses of Bayesian networks as tools for synthesis of systematic reviews of complex interventions

Bayesian networks (BNs) are tools for representing expert knowledge or evidence. They are especially useful for synthesising evidence or belief concerning a complex intervention, assessing the sensitivity of outcomes to different situations or contextual frameworks and framing decision problems that involve alternative types of intervention. Bayesian networks are useful extensions to logic maps when initiating a review or to facilitate synthesis and bridge the gap between evidence acquisition and decision-making. Formal elicitation techniques allow development of BNs on the basis of expert opinion. Such applications are useful alternatives to 'empty' reviews, which identify knowledge gaps but fail to support decision-making. Where review evidence exists, it can inform the development of a BN. We illustrate the construction of a BN using a motivating example that demonstrates how BNs can ensure coherence, transparently structure the problem addressed by a complex intervention and assess sensitivity to context, all of which are critical components of robust reviews of complex interventions. We suggest that BNs should be utilised to routinely synthesise reviews of complex interventions or empty reviews where decisions must be made despite poor evidence.

The cost effectiveness of pharmacological treatments for generalized anxiety disorder

BACKGROUND

Generalized anxiety disorder (GAD) is one of the most prevalent anxiety disorders, with important implications for patients and healthcare resources. However, few economic evaluations of pharmacological treatments for GAD have been published to date, and those available have assessed only a limited number of drugs.

OBJECTIVE

To assess the cost effectiveness of pharmacological interventions for patients with GAD in the UK.

METHODS

A decision-analytic model in the form of a decision tree was constructed to compare the costs and QALYs of six drugs used as first-line pharmacological treatments in people with GAD (duloxetine, escitalopram, paroxetine, pregabalin, sertraline and venlafaxine extended release [XL]) and 'no pharmacological treatment'. The analysis adopted the perspective of the NHS and Personal Social Services (PSS) in the UK. Efficacy data were derived from a systematic literature review of double-blind, randomized controlled trials and were synthesized using network meta-analytic techniques. Two network meta-analyses were undertaken to assess the comparative efficacy (expressed by response rates) and tolerability (expressed by rates of discontinuation due to intolerable side effects) of the six drugs and no treatment in the study population. Cost data were derived from published literature and national sources, supplemented by expert opinion. The price year was 2011. Probabilistic sensitivity analysis was conducted to evaluate the underlying uncertainty of the model input parameters.

RESULTS

Sertraline was the best drug in limiting discontinuation due to side effects and the second best drug in achieving response in patients not discontinuing treatment due to side effects. It also resulted in the lowest costs and highest number of QALYs among all treatment options assessed. Its probability of being the most cost-effective drug reached 75 % at a willingness-to-pay threshold of £20,000 per extra QALY gained.

CONCLUSION

Sertraline appears to be the most cost-effective drug in the treatment of patients with GAD. However, this finding is based on limited evidence for sertraline (two published trials). Sertraline is not licensed for the treatment of GAD in the UK, but is commonly used by primary care practitioners for the treatment of depression and mixed depression and anxiety.

Clinical recognition and recording of alcohol disorders by clinicians in primary and secondary care: meta-analysis

BACKGROUND

Clinicians have considerable difficulty identifying and helping those people with alcohol problems but no previous study has looked at this systematically.

AIMS

To determine clinicians' ability to routinely identify broadly defined alcohol problems.

METHOD

Data were extracted and rated by two authors, according to PRISMA standard and QUADAS criteria. Studies that examined the diagnostic accuracy of clinicians' opinion regarding the presence of alcohol problems as well as their written notation were evaluated.

RESULTS

A comprehensive search identified 48 studies that looked at the routine ability of clinicians to identify alcohol problems (12 in primary care, 31 in general hospitals and 5 in psychiatric settings). A total of 39 examined alcohol use disorder, 5 alcohol dependence and 4 intoxication. We separated studies into those using self-report and those using interview. The diagnostic sensitivity of primary care physicians (general practitioners) in the identification of alcohol use disorder was 41.7% (95% CI 23.0-61.7) but alcohol problems were recorded correctly in only 27.3% (95% CI 16.9-39.1) of primary care records. Hospital staff identified 52.4% (95% CI 35.9-68.7) of cases and made correct notations in 37.2% (95% CI 28.4-46.4) of case notes. Mental health professionals were able to correctly identify alcohol use disorder in 54.7% (95% CI 16.8-89.6) of cases. There were limited data regarding alcohol dependency and intoxication. Hospital staff were able to detect 41.7% (95% CI 16.5-69.5) of people with alcohol dependency and 89.8% (95% CI 70.4-99.4) of those acutely intoxicated. Specificity data were sparse.

CONCLUSIONS

Clinicians may consider simple screening methods such as self-report tools rather than relying on unassisted clinical judgement but the added value of screening over and above clinical diagnosis remains unclear.

Diagnostic validity of the Hospital Anxiety and Depression Scale (HADS) in cancer and palliative settings: a meta-analysis

OBJECTIVE

To examine the validity of the Hospital Anxiety and Depression Scale (HADS) in the identification of psychiatric complications of cancer, as defined by a robust criterion standard.

METHODS

50 analyses tested the depression subscale (HADS-D), anxiety subscale (HADS-A) or combined scales (HADS-T) against syndromal (clinical) depression (n=22), syndromal anxiety (n=4) or any mental ill health/distress (n=24), all defined by semi-structured psychiatric interview.

RESULTS

The HADS and its subscales had both strengths and limitations. Overall it appeared to perform marginally better in non-palliative cancer settings. Specific findings for each subscale were as follows. In the identification of depression the HADS-T, HADS-D and HADS-A had a pooled sensitivity and specificity of 82.0%, 77.0%; 71.6%, 82.6% and 80.5%, 77.8%, respectively. All versions performed poorly in case-finding but well in a screening capacity. For anxiety there were no HADS-D studies. The HADS-T and HADS-A had a pooled sensitivity and specificity of 83.9%, 69.9% and 48.7%, 78.7%. They were poor at case-finding but good as screening instruments. For distress (any mental ill health) the HADS-T, HADS-D and HADS-A had a pooled sensitivity and specificity of 72.8%, 80.6%; 75.7%, 66.3% and 65.7%, 71.3%, respectively. When screening for distress and anxiety the HADS-T was the optimal subscale.

CONCLUSION

For the identification of depression, anxiety or distress in cancer settings, the HADS (including subscales) is not recommended as a case-finding instrument but it may, subject to concerns about its length, be a suitable addition to screening programme.

Diagnostic validity and added value of the Geriatric Depression Scale for depression in primary care: a meta-analysis of GDS30 and GDS15

BACKGROUND

The Geriatric Depression Scale (GDS) has been evaluated in hospital settings but its validity and added value in primary care is uncertain. We therefore conducted a meta-analysis analysing the diagnostic accuracy, clinical utility and added value of the GDS in primary care.

METHODS

A comprehensive search identified 69 studies that measured the diagnostic validity of the GDS against a semi-structured psychiatric interview and of these 17 analyses (in 14 publications) took place in primary care. Seven studies examined the GDS(30) and 10 studies examined the GDS(15). Heterogeneity was moderate to high, therefore random effects meta-analysis was used.

RESULTS

Diagnostic accuracy of the GDS(30) after meta-analytic weighting was given by a sensitivity of 77.4% (95% CI=66.3% to 86.8%) and a specificity=65.4% (95% CI=44.2% to 83.8%). For the GDS(15) the sensitivity was 81.3% (95% CI=77.2% to 85.2%) and specificity=78.4% (95% CI=71.2% to 84.8%). The fraction correctly identified (also known as efficiency) by the GDS(15) was significantly higher than the GDS(30) (77.6% vs 71.2%, Chi(2)=24.8 P<0.0001). The clinical utility of both the GDS(30) and GDS(15) was "poor" for case-finding (UI+ 0.29, UI+ 0.32 respectively). However the GDS(15) was rated as "good" for screening (UI- 0.75) whereas the GDS(30) was "adequate" (UI- 0.60). Concerning added value, when identification using the GDS was compared with general practitioners' ability to diagnose late-life depressions unassisted by tools, at a prevalence of 15% the GDS(30) had no added benefit whereas the GDS(15) helped identify an additional 4 cases per 100 primary care attendees and also helped rule-out an additional 4 non-cases per 100 attendees. Thus we estimate the potential gain of the GDS(15) in primary care to be 8% over unassisted clinical detection but at a cost of 3-4 minutes of extra time per appointment.

CONCLUSION

The GDS yields potential added value in primary care. We recommend the GDS(15) but not the GDS(30) in the diagnosis of late-life depression in primary care.

Factors affecting nonmedical participants' allocation of scarce medical resources

This study was designed to determine the factors that affect nonmedical participants' judgments in constructing a ranked waiting list for kidney patients requiring dialysis. Participants (N=167) were given a questionnaire that provided minimal demographic data about 16 hypothetical patients. Participants were requested to rank patients in order of priority for treatment. Each participant's personal demographic details were also obtained. Patients differed on four dimensions: gender, income, alcohol consumption, and religious beliefs, yielding a 2x2x2x2 design. The participants favoured for treatment included females over males, "poor" over "rich," nondrinkers over drinkers, and Christians over atheists. Results are discussed in terms of establishing democratic criteria and informing medical personnel on explicit factors which may affect their decision making, thus guarding against biases in judgment.