Computational pathology in cancer diagnosis, prognosis, and prediction - present day and prospects

Computational pathology refers to applying deep learning techniques and algorithms to analyse and interpret histopathology images. Advances in artificial intelligence (AI) have led to an explosion in innovation in computational pathology, ranging from the prospect of automation of routine diagnostic tasks to the discovery of new prognostic and predictive biomarkers from tissue morphology. Despite the promising potential of computational pathology, its integration in clinical settings has been limited by a range of obstacles including operational, technical, regulatory, ethical, financial, and cultural challenges. Here, we focus on the pathologists' perspective of computational pathology: we map its current translational research landscape, evaluate its clinical utility, and address the more common challenges slowing clinical adoption and implementation. We conclude by describing contemporary approaches to drive forward these techniques. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Abstract PS8-22: Augmentation of a minimal multidisciplinary tumor board with clinical decision support to triage breast cancer patients in the UK

BackgroundAll UK cancer patients undergo required assessments by a full Multidisciplinary Tumor Board (fMTB) at key treatment decision points, placing a resource burden on the healthcare system. Watson for Oncology (WfO) is a decision-support system that presents therapeutic options to cancer-treating clinicians. This study is an initial phase of an evaluation at Guys and St. Thomas’ NHS Hospital (GSTT), designed to explore the extent to which WfO can be used by the fMTB to triage less complex patient cases and ultimately reduce workload and time pressures currently experienced by fMTBs. We conducted a concordance study with two minimal MTB teams (mMTB) for Stage I-III breast cancer patients.

MethodsBreast cancer cases (N=63) treated from 2017-2018 at GSTT were evaluated by 2 independent mMTBs, blinded to each other and previous fMTB decisions rendered prior to this study. Each mMTB consisted of a senior medical oncologist and surgeon; GSTT’s 12+ member fMTB is comprised of oncologists, surgeons, radiologists, pathologists and others. mMTBs were shown options that were either listed as ‘recommended’ or ‘for consideration’ by WfO and given the opportunity to revise prior decisions. The combined 4-person minimal MTB (cmMTB) consisting of both 2-person mMTBs provided a current consensus best-practice plan and systemic therapy recommendations for discordant cases. We evaluated the concordance of WfO’s systemic therapeutic recommendations and mMTBs, as well as concordance with the cmMTB. Previous decisions by the fMDTB were also compared to decisions by the cmMTB. Univariate logistic regression explored characteristics predictive of concordance with the cmMTB.

ResultsFor treatment plans, WFO’s therapeutic options had higher concordance with cmMTB decisions than either mMTB alone (concordance 93.7% vs. 92.1%) or the previous decisions by the fMTB (87.3). For systemic therapy decisions, the WfO-cmMDTB concordance was 70.2%; however, adjusting for non-NICE approved drugs and the common practice of Carboplatin use in the UK, concordance increased to 91.5%. Previous decisions by the fMTB had the lowest concordance with the cmMTB (87.3%). Adjusting for the UK-practice related use of Carboplatin, WfO had slightly higher concordance with cmMTB systemic therapy decisions than either mMTB alone (89.4% and 87.2%). Univariate analysis with this limited sample revealed non-significant trends in association between mMTB’s concordance with WfO and stage of cN at diagnosis, HER2 status, tumor location and grade. For example, mMTBs concordance with WfO tended to improve when tumor grade was high. Non-significant trends were also identified in the association between WfO-treatment concordance and tumor location, where treatment concordance increased with medial tumor location.

ConclusionIn this small cohort study, a clinical decision-support tool demonstrated better agreement with UK best practice treatment than a 2-person mMTB and may have a role in triaging breast cancer cases in the UK.

Citation Format: Hartmut Kristeleit, Martha Martin, Christina Karampera, Rezzan Hekmat, Bertha IntHout, Ashutosh Kothari, Majid Kazmi, Amanda Clery, Yanzhong Wang, Bolaji Coker, Winnie Felix, Anita Preininger, Suwei Wang, Roy Vergis, Tom Eggebraaten, Christopher Gloe, Irene Dankwa-Mullan, Gretchen Jackson, Anna Rigg, Danny Ruta. Augmentation of a minimal multidisciplinary tumor board with clinical decision support to triage breast cancer patients in the UK [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS8-22.

Portuguese translation and validation of the Patient Generated Index (PGI) instrument for patients with Chronic Obstructive Pulmonary Disease: individualized quality of life assessment

Objective To translate, adapt and validate the Patient Generated Index (PGI) for Brazilians with chronic obstructive pulmonary disease (COPD). Methods 50 volunteers with COPD, mostly men (74%), with 73.1 ± 8.9 years of age, FEV1 of 52.3 ± 14.5% of predicted and FEV1 / FVC of 56.2 ± 8.6% of predicted responded to PGI, to the Saint George Respiratory Questionnaire (SGRQ) and to perform Glittre Activities of Daily Living test (Glittre ADL). After 1-2 weeks, PGI was again applied for the analysis of relative and absolute reliability. Results The translation occurred without changes in the questionnaire. The score obtained in PGI had weak correlation with the SGRQ total score (r = -0.44, p <0.001) and with the impact domain (r = -0.40, p <0.05), presented a moderate correlation with the symptoms domain of the SGRQ (r = -0.55, p <0.001) and weak correlation with the activity domain (r = -0.31, p <0.05). A weak correlation was observed between PGI and Glittre ADL (r = -0.30; p <0.05). It was observed high reliability among the measures of PGI (ICCr = 0.94). Conclusion This study shows that the Brazilian version of PGI is a reliable and valid instrument to measure health-related quality of life (HRQL) in patients with COPD. It is a new and individualized form of evaluation of COPD patient-centered quality of life.

The Daily Mile as a public health intervention: a rapid ethnographic assessment of uptake and implementation in South London, UK

Background

Existing evidence identifies health benefits for children of additional daily physical activity (PA) on a range of cardiovascular and metabolic outcomes. The Daily Mile (TDM) is a popular scheme designed to increase children’s PA within the school day. Emerging evidence indicates that participation in TDM can increase children’s PA, reduce sedentarism and reduce skinfold measures. However, little is known about the potential effects of TDM as a public health intervention, and the benefits and disbenefits that might flow from wider implementation in ‘real world’ settings.

Methods

We aimed to identify how TDM is being implemented in a naturalistic setting, and what implications this has for its potential impact on population health. We undertook a rapid ethnographic assessment of uptake and implementation in Lewisham, south London. Data included interviews (n = 22) and focus groups (n = 11) with stakeholders; observations of implementation in 12 classes; and analysis of routine data sources to identify school level factors associated with uptake.

Results

Of the 69 primary schools in one borough, 33 (48%) had adopted TDM by September 2018. There were no significant differences between adopters and non-adopters in mean school population size (means 377 vs 397, P = 0.70), mean percentage of children eligible for free school meals (16.2 vs 14.3%, P = 0.39), or mean percentage of children from Black and Minority Ethnic populations (76.3 vs 78.2%, P = 0.41). Addressing obesity was a key incentive for adoption, although a range of health and educational benefits were also hypothesised to accrue from participation. Mapping TDM to the TIDierR-PHP checklist to describe the intervention in practice identified that considerable adaption happened at the level of borough, school, class and pupil. Population health effects are likely to be influenced by the interaction of intervention and context at each of these levels.

Conclusions

Examining TDM in ‘real world’ settings surfaces adaptions and variations in implementation. This has implications for the likely effects of TDM, and points more broadly to an urgent need for more appropriate methods for evaluating public health impact and implementation in complex contexts.

Electronic supplementary material

The online version of this article (10.1186/s12889-019-7511-9) contains supplementary material, which is available to authorized users.

Priority setting in healthcare: towards guidelines for the program budgeting and marginal analysis framework

Economists' approaches to priority setting focus on the principles of opportunity cost, marginal analysis and choice under scarcity. These approaches are based on the premise that it is possible to design a rational priority setting system that will produce legitimate changes in resource allocation. However, beyond issuing guidance at the national level, economic approaches to priority setting have had only a moderate impact in practice. In particular, local health service organizations - such as health authorities, health maintenance organizations, hospitals and healthcare trusts - have had difficulty implementing evidence from economic appraisals. Yet, in the context of making decisions between competing claims on scarce health service resources, economic tools and thinking have much to offer. The purpose of this article is to describe and discuss ten evidence-based guidelines for the successful design and implementation of a program budgeting and marginal analysis (PBMA) priority setting exercise. PBMA is a framework that explicitly recognizes the need to balance pragmatic and ethical considerations with economic rationality when making resource allocation decisions. While the ten guidelines are drawn from the PBMA framework, they may be generalized across a range of economic approaches to priority setting.

Priority setting in practice: what is the best way to compare costs and benefits?

Prioritizing candidates for health-care expenditure using cost per Quality-Adjusted Life Year (QALY) is a helpful but insufficient means of ranking alternative uses for scarce health-care funds at the local level. This is because QALYs do not by themselves capture all criteria decision makers need to take into account. Other criteria such as reducing inequalities, meeting national and local priorities and public acceptability also feature in the decision maker's utility function. Programme budgeting and marginal analysis (PBMA) is an established framework for systematic priority setting in which a 'weighted benefit score' for each option is calculated based on all relevant decision-making criteria. Ranking options as a ratio of cost to benefit is desirable and necessary to ensure efficiency. In this paper we review a number of approaches to scoring costs and benefits of options in a PBMA context. Several approaches rank by benefit score alone, rather than efficiency (cost per unit of benefit). Of those that do rank by efficiency, we discuss the benefits and drawbacks. The optimal approach is far from clear, with each technique having its own strengths and weaknesses. A deliberative approach using summaries of costs and benefits of options as a basis for discussion may be preferable.

'Translation is not enough': using the Global Person Generated Index (GPGI) to assess individual quality of life in Bangladesh, Thailand, and Ethiopia

Currently few subjective measures of Quality of Life (QoL) are available for use in developing countries, which limits their theoretical, methodological, and practical contribution (for example, exploring the relationship between economic development and QoL, and ensuring effective and equitable service provision). One reason for this is the difficulty of ensuring that translated measures preserve conceptual, item, semantic, operational, measurement; and functional equivalence (Herdman, M., Fox-Rushby, J., & Badia, X. (1998). Quality of Life Research, 7, 331), which is illustrated by an account of the translation, pre-piloting, and administration of a new individualised QoL measure, the Global Person Generated Index or 'GPGI'. The GPGI is based on the widely used Patient Generated Index (Ruta, Camfield, & Martin, (2004) Quality of Life Research, 13, 1545.) and offers many of the advantages of the participatory approaches commonly used in developing countries, with added methodological rigour, and quantitative outcomes. It was successfully validated in Bangladesh, Thailand, and Ethiopia, using quantitative and qualitative methods--open-ended, semi-structured interviews (SSIs), conducted immediately post-administration. Both the measure and method of 'qualitative validation' described later in the paper offer an exciting alternative for future researchers and practitioners in this field. The quantitative results suggest the GPGI shows cultural sensitivity, and is able to capture both the areas that are important to respondents, and aspects of life one would expect to impact on QoL in developing countries. There were strong correlation between scores from the GPGI and SSIs for the area of health, and moderate correlations for 'material wellbeing' (MWB)('Material wellbeing' refers to respondents' perceptions of their achievement in the areas of farming, debt reduction, assets, crops, livestock, job, land, property, and agriculture) and children. Weak to moderate correlations were observed between the Satisfaction with Life Scale and the GPGI; however, the highest coefficient was between the GPGI and the most conceptually similar item. Statistically significant differences were seen in GPGI scores between rich and poor, urban and rural respondents, and different countries. Health and material wellbeing scores, derived from the SSIs, also showed a linear relationship with GPGI scores, with a suggestion of curvilinearity at the higher levels, as predicted by a general QoL causal model. In conclusion, the GPGI has great potential for use in this area, especially when supported by extensive interviewer training, and supplemented with a cognitive appraisal schedule.

Twelve years–experience with the Patient Generated Index (PGI) of quality of life: a graded structured review

The Patient Generated Index (PGI) is an individualised quality of life (QoL) measure that has been in use since 1994. Various adaptations have been made to suit a variety of client groups. The PGI's psychometric properties have been studied but their review is necessary to inform instrument choice. This article provides a structured review, using grading criteria adapted from those developed to aid outcome measure selection for use with older people. These criteria grade quality of evidence and strength of findings for psychometric validity, providing a useful model for future reviews. All published articles providing data addressing validity, reliability and/or responsiveness were included in the review. Eighteen relevant articles were identified and analysed using the grading criteria. Variable results and quality of investigation were seen. Generally the measure was found to be adequately reliable for group comparisons. The PGI appeared valid but evidence for responsiveness was unclear. Those versions of the measure using fewer points in their Likert scales may have higher reliability. Cognitions involved in QoL judgements remain little understood and investigations of psychometric properties may be enhanced by examination of appraisal processes.

Using economics to set pragmatic and ethical priorities

Doctors and managers have to make tough decisions about what services to provide from their budgets. Economic approaches can help, but they also need to take into account the practical and ethical challenges faced by healthcare professionals

Should the NHS follow the American way?

Managed care and patient choice have many good points, but the NHS needs to adapt US methods if it is to be efficient, free, and fair

Programme budgeting and marginal analysis: bridging the divide between doctors and managers

Recent NHS reforms give doctors increased responsibility for efficient and fair use of resources. Programme budgeting and marginal analysis is one way to ensure the views of all stakeholders are properly represented

Implementation strategies for a Scottish national epilepsy guideline in primary care: results of the Tayside Implementation of Guidelines in Epilepsy Randomized (TIGER) trial

PURPOSE

To determine the effectiveness of two dissemination and implementation strategies to implement a national guideline for epilepsy management in primary care settings.

METHODS

Three-arm cluster-randomized controlled trial. The participants were general practitioners from 68 practices in Tayside, Scotland, and 1,133 of their patients with self-reported epilepsy treated with antiepileptic medications (AEDs). Practices were randomized blind to a control, intermediate, or intensive intervention.

CONTROL

Postal dissemination of a nationally developed clinical guideline. Intermediate intervention: Postal dissemination of the guideline supported by interactive, accredited workshops, and dedicated, structured protocol documents. Intensive intervention: Intermediate intervention plus a nurse specialist who supported and educated practices in the establishment of epilepsy review clinics. The primary outcome was the SF-36 health-related quality-of-life instrument. Secondary measures were a battery of prevalidated epilepsy-specific quality-of-life instruments. These were administered at baseline and after the intervention phase. Process of care was assessed by case-note review on number of review meetings and counseling sessions for epilepsy before and after the interventions.

RESULTS

None of the intervention groups showed any change in the primary or secondary outcome measures or process-of-care measures.

CONCLUSIONS

None of the intervention strategies led to improvements in patient quality of life or quality of epilepsy care. Further research is needed to discover why the interventions failed, to identify barriers to adoption of guidelines, and to develop strategies that might improve implementation and uptake in the future.

Pilot study: quality of life assessment of postnatal fatigue and other physical morbidity

Postnatal morbidity is increasingly recognized, but standard assessments may not capture what is most important to the woman with such morbidity in terms of her quality of life. The Mother-Generated Index (MGI) is a proposed postnatal quality-of-life instrument which allows the mother to determine both content and scoring. In this pilot study we found that although a degree of psychological and physical morbidity (including tiredness) is common, and may be very significant, for most women these factors are low-grade, and other aspects of their lives are more important. A quality-of-life approach allows the mother to determine her own postnatal assessment, and encourages practitioners to view her more holistically.

Postnatal quality of life: a pilot study using the Mother-Generated Index

BACKGROUND

Although postnatal morbidity has been well documented in recent years, postnatal quality of life has not been addressed. A newly derived subjective measurement of postnatal quality of life (the Mother-Generated Index) combines a quantitative and qualitative evaluation.

AIMS

This part of our pilot study aimed to compare the aspects of their lives nominated by women with low and high quality of life (Primary Index) scores, and to examine the respective importance of these areas.

METHODS

The Mother-Generated Index was tested using the Edinburgh Postnatal Depression Scale, Short Form 12, and an established maternal and neonatal physical morbidity index as validators. Four health visitors administered these at 6-8 weeks and 8 months postpartum to 103 women by structured face-to-face interviews between June 2000 and March 2001. Data were entered into Epi-Info, and exported to Microsoft Excel and SPSS for analysis.

RESULTS

A wide variety of quality of life aspects were reported, including emotional, social and financial concerns. Tiredness was prevalent in all groups, but other physical problems were rare at 8 months. Mothers with low quality of life (Primary Index) scores at 6-8 weeks and 8 months commonly reported having less personal time. Low scoring areas, which health professionals might consider in greatest need of attention, were often not the ones mothers deemed most important.

LIMITATIONS

The study involved only 103 participants, and did not assess the degree of support experienced by the mothers.

CONCLUSIONS

The Mother-Generated Index helps mothers to identify the areas of their lives which are of most concern to them. This pilot suggests that mothers with high and low quality of life scores have markedly divergent experiences.

Postnatal quality of life assessment: introducing the mother-generated index

BACKGROUND

The extent of postnatal morbidity has become increasingly apparent over the last 15 years, but currently no tool is available that measures postnatal quality of life. This pilot study introduces a subjective tool, the Mother-Generated Index, which assesses the woman's quality of life and identifies those aspects that are of most concern to her.

METHODS

The Mother-Generated Index was administered by structured interview to 60 participants at 6 to 8 weeks and to 43 participants at 8 months postpartum. Validation was sought through concurrent use of the Edinburgh Postnatal Depression Scale, the SF12, and two indexes related to maternal and neonatal physical morbidity. The Mother-Generated Index gives a primary index (quality of life) score, which is reported here, and a secondary index, which identifies the areas considered most important by the mother.

RESULTS

The primary index was more sensitive at 8 months. The highest and lowest quartile scores were compared. Statistically significant differences in were found in the mothers' Edinburgh Postnatal Depression scores at 6 to 8 weeks, and in their Edinburgh Postnatal Depression and SF12 mental component scores and their physical morbidity index at 8 months. Although physical problems were only a small feature at 8 months, social and psychological issues were prominent in both groups. Age, parity, and mode of delivery had no significant effect on the women's scores or the areas they identified as most important.

CONCLUSION

Quality of life of is a complex and personal area, affected by many different aspects of health and well-being. From this pilot study the primary index appears to be a useful step in assessing a mother's quality of life. It identifies which areas of her life are most important to her, and allows her to indicate where she would like to see improvements.

Hemispheric-synchronisation during anaesthesia: a double-blind randomised trial using audiotapes for intra-operative nociception control

The possible antinociceptive effect of hemispheric-synchronised sounds, classical music and blank tape were investigated in patients undergoing surgery under general anaesthesia. The study was performed on 76 patients, ASA 1 or 2, aged 18-75 years using a double-blind randomised design. Each of the three tapes was allocated to the patients according to a computer-generated random number table. General anaesthesia was standardised and consisted of propofol, nitrous oxide 66%/oxygen 33%, isoflurane and fentanyl. Patients breathed spontaneously through a laryngeal mask and the end-tidal isoflurane concentration was maintained near to its minimum alveolar concentration value of 1.2%. Fentanyl was given intravenously sufficient to keep the intra-operative heart rate and arterial blood pressure within 20% of pre-operative baseline values and the fentanyl requirements were used as a measure of nociception control. Patients to whom hemispheric-synchronised sounds were played under general anaesthesia required significantly less fentanyl compared with patients listening to classical music or blank tape (mean values: 28 microgram, 124 microgram and 126 microg, respectively) (p < 0.001). This difference remained significant when regression analysis was used to control for the effects of age and sex.

Audit of antibiotic prophylaxis for surgical patients in three hospital trusts in Tayside. Tayside Area Clinical Audit Commitee

OBJECTIVE

To compare achievement of previously agreed standards for administration of antibiotic prophylaxis for surgical patients in the three acute trusts in Tayside.

SETTING

Angus, Dundee Teaching Hospitals and Perth & Kinross Trusts.

MAIN OUTCOME MEASURES

Administration of antibiotic prophylaxis and achievement of three general standards of administration.

STUDY DESIGN

Prospective case note audit.

RESULTS

In total, 341 operations were audited. The range of operations performed in each trust was different and the achievement of standards was sensitive to case mix. For example, prophylaxis was given to 82% of all eligible orthopaedic patients but only 64% of eligible patients in obstetrics & gynaecology. Comparison between trusts was therefore confined to six procedures (166 operations) which were performed regularly in all three. There were significant differences between trusts in the rate of administration to patients undergoing operations for which prophylaxis was indicated (from 84% [95% CI 75-92] to 93% [CI 89-98]) and in the proportion of patients in whom prophylaxis was continued for less than 24 hours (from 78% [CI 68-89] to 97% [CI 93-100]). Administration of prophylaxis within two hours of surgery was achieved following more than 95% of operations in all three trusts. However, second doses were given to only five of 44 (11%) of patients whose operations lasted more than two hours.

CONCLUSIONS

Achievement of standards in all three trusts was good in comparison with recently published audits from other UK and European centres but there was still room for improvement, particularly in administration of second doses for prolonged operations. Regular audit of prophylaxis administration and duration should be implemented. Comparisons between trusts or units should be based on a common set of operations.

Seat restraint use, previous driving history, and non-fatal injury: quantifying the risks

AIMS

To quantify the increased risk of non-fatal injury when children travel unrestrained in a car, and to identify other preventable risk factors.

METHODS

Case-control study of 78 children presenting to an accident and emergency (A&E) department having sustained an injury while travelling in a car, and 97 children attending an A&E outpatient clinic with conditions unrelated to road traffic accidents.

RESULTS

Seat restraint was associated with a 93% lower risk of child accident injury. A driver with points on the licence was over five times more likely to have had an accident resulting in child injury than a driver without points. Child accident injury was also associated with the driver's accident history.

CONCLUSIONS

These data allow the effect of achieving new target levels of seat restraint use to be estimated. Strategies aimed at reducing the risk of further accident among drivers with a history of accident may have a beneficial impact on childhood accident injuries.

Measuring health-related quality of life in rheumatoid arthritis: validity, responsiveness and reliability of EuroQol (EQ-5D)

The EuroQol (EQ-5D) generic health index comprises a five-part questionnaire and a visual analogue self-rating scale. The questionnaire may be used as a health index to calculate a 'utility' value or as a health profile. The validity, reliability and responsiveness of EQ-5D were tested in 233 patients with rheumatoid arthritis stratified by functional class. EQ-5D demonstrated moderate to high correlations with measures of impairment and high correlations with disability measures. Stepwise regression models showed that EQ-5D utility values and visual analogue scores were explained best as a function of pain, disability, disease activity and mood (R2 approximately 70%), although other variables (side-effects, years of education) were required to explain the visual analogue scores. The EQ-5D health index and visual analogue scale are more responsive than any of the other measures, except pain and doctor-assessed disease activity. The reliability of the EQ-5D index and EQ-5D visual analogue scale is as good or better than that of all other instruments except the Health Assessment Questionnaire. Some patients with severe long-standing disease had health states which attracted utility values below zero, i.e. from a societal perspective they were regarded as being in states 'worse than death'. The practical and ethical implications of these utility valuations are discussed, and at present the utility values should be used and interpreted with caution. With this caveat, EQ-5D is simple to use, valid, responsive to change and sufficiently reliable for group comparisons. It is of potential use as an outcome measure in clinical trials, audit and health economic studies, but further work is required on its performance in other clinical contexts and on the interpretation of the utility values.

Feasibility of monitoring patient based health outcomes in a routine hospital setting

OBJECTIVE

To assess the feasibility of monitoring health outcomes in a routine hospital setting and the value of feedback of outcomes data to clinicians by using the SF 36 health survey questionnaire.

DESIGN

Administration of the questionnaire at baseline and three months, with analysis and interpretation of health status data after adjustments for sociodemographic variables and in conjunction with clinical data. Exploration of usefulness of outcomes data to clinicians through feedback discussion sessions and by an evaluation questionnaire.

SETTING

One gastroenterology outpatient department in Aberdeen Royal Hospitals Trust, Scotland.

PATIENTS

All (573) patients attending the department during one month (April 1993).

MAIN MEASURES

Ability to obtain patient based outcomes data and requisite clinical information and feed it back to the clinicians in a useful and accessible form.

RESULTS

Questionnaires were completed by 542 (95%) patients at baseline and 450 (87%) patients at follow up. Baseline health status data and health outcomes data for the eight different aspects of health were analysed for individual patients, key groups of patients, and the total recruited patient population. Significant differences were shown between patients and the general population and between different groups of patients, and in health status over time. After adjustment for differences in sociodemography and main diagnosis patients with particularly poor scores were identified and discussed. Clinicians judged that this type of assessment could be useful for individual patients if the results were available at the time of consultation or for a well defined group of patients if used as part of a clinical trial.

CONCLUSIONS

Monitoring routine outcomes is feasible and instruments to achieve this, such as the SF 36 questionnaire, have potential value in an outpatient setting.

IMPLICATIONS

If data on outcomes are to provide a basis for clinical and managerial decision making, information systems will be required to collect, analyse, interpret, and feed it back regularly and in good time.

A prospective study of non-fatal childhood road traffic accidents: what can seat restraint achieve?

A prospective study of children attending an Accident and Emergency Department was carried out at the Royal Aberdeen Children's Hospital, to determine the association between injury risk and seat restraint usage, and to estimate the likely impact of seat restraint usage at various target levels on the incidence of non-fatal injuries. The subjects were all 91 cases of non-fatal childhood road traffic accidents attending the casualty department between December 1989 and November 1990. The main outcome measures were relative risks of three grades of injury severity and four types of non-fatal injury, population attributable risk fractions and estimated likely reduction in incidence of non-fatal injury at three set target levels of seat restraint usage. It was found that 42 (46 per cent) children presenting to the A&E Department were unrestrained; 78 (85.7 per cent) children sustained some injury. The relative risk for children travelling unrestrained in a car for all injuries combined is 1.7 (95 per cent confidence interval (CI) 1.17-2.45), the relative risk for head injury is 3.13 (1.78-5.51) and for face injury 3.03 (1.44-6.37). The risk of sustaining any moderate or worse injury was 3.25 (1.05-10.07). It is estimated that 24.4 per cent (minimum 5.5 per cent; maximum 45 per cent) of all non-fatal injuries sustained by a child car passenger can be prevented if all children are restrained in a child safety seat or seat restraint; 49.5 per cent of head injuries (minimum 43.4 per cent; maximum 71.9 per cent) and 48.4 per cent of face injuries (minimum 21.9 per cent; maximum 75.3 per cent) are preventable.(ABSTRACT TRUNCATED AT 250 WORDS)