Introduction Twitter and other social media are increasingly used by patients to discuss their experiences of healthcare. Social media might provide a new way for health services to listen to the voices of patients and improve their services. Little is known about how patients are communicating with hospitals via this route, and whether there is any association with traditional measures of patient experience such as surveys. Methods We recorded tweets aimed at all acute hospital trusts with Twitter accounts in England for one year from April 2012. We performed a qualitative content analysis of a random sample of 1000 tweets, to see what information they contained about care quality. Using natural language processing techniques, we calculated the sentiment of all the tweets towards hospital. We compared twitter sentiment to patient experience measured by traditional survey at the hospital level, using Spearman's rank correlation coefficient. Results We collected 187,000 tweets over one year. The mean number of tweets per trust was 2499. 9.8% of tweets were related to quality of care care – and most of these related to patients' experience of interactions with staff. We found no correlation between the sentiment of tweets about hospitals and patient experience measure by traditional survey methodology (Rho=0.08, p=0.56). Discussion Although social media are increasingly used by both the public and healthcare professionals to communicate, caution should be taken in using social media data to measure care quality. The information contained within tweets was able to provide valuable individual insights about some patients' experiences of care, however the views expressed appeared less likely to be representative of the experiences of the wider population receiving care. Declaration of competing interests This work has been funded by the Commonwealth Fund.

Background Twitter is increasingly being used by patients to comment on their experience of healthcare. This may provide information for understanding the quality of healthcare providers and improving services.Objective To examine whether tweets sent to hospitals in the English National Health Service contain information about quality of care. To compare sentiment on Twitter about hospitals with established survey measures of patient experience and standardised mortality rates.Design A mixed methods study including a quantitative analysis of all 198 499 tweets sent to English hospitals over a year and a qualitative directed content analysis of 1000 random tweets. Twitter sentiment and conventional quality metrics were compared using Spearman's rank correlation coefficient.Key results 11% of tweets to hospitals contained information about care quality, with the most frequent topic being patient experience (8%). Comments on effectiveness or safety of care were present, but less common (3%). 77% of tweets about care quality were positive in tone. Other topics mentioned in tweets included messages of support to patients, fundraising activity, self-promotion and dissemination of health information. No associations were observed between Twitter sentiment and conventional quality metrics.Conclusions Only a small proportion of tweets directed at hospitals discuss quality of care and there was no clear relationship between Twitter sentiment and other measures of quality, potentially limiting Twitter as a medium for quality monitoring. However, tweets did contain information useful to target quality improvement activity. Recent enthusiasm by policy makers to use social media as a quality monitoring and improvement tool needs to be carefully considered and subjected to formal evaluation.

Recent years have seen increasing interest in patient-centred care and calls to focus on improving the patient experience. At the same time, a growing number of patients are using the internet to describe their experiences of healthcare. We believe the increasing availability of patients' accounts of their care on blogs, social networks, Twitter and hospital review sites presents an intriguing opportunity to advance the patient-centred care agenda and provide novel quality of care data. We describe this concept as a 'cloud of patient experience'. In this commentary, we outline the ways in which the collection and aggregation of patients' descriptions of their experiences on the internet could be used to detect poor clinical care. Over time, such an approach could also identify excellence and allow it to be built on. We suggest using the techniques of natural language processing and sentiment analysis to transform unstructured descriptions of patient experience on the internet into usable measures of healthcare performance. We consider the various sources of information that could be used, the limitations of the approach and discuss whether these new techniques could detect poor performance before conventional measures of healthcare quality.

Background: There are large amounts of unstructured, free-text information about quality of health care available on the Internet in blogs, social networks, and on physician rating websites that are not captured in a systematic way. New analytical techniques, such as sentiment analysis, may allow us to understand and use this information more effectively to improve the quality of health care.Objective: We attempted to use machine learning to understand patients' unstructured comments about their care. We used sentiment analysis techniques to categorize online free-text comments by patients as either positive or negative descriptions of their health care. We tried to automatically predict whether a patient would recommend a hospital, whether the hospital was clean, and whether they were treated with dignity from their free-text description, compared to the patient's own quantitative rating of their care.Methods: We applied machine learning techniques to all 6412 online comments about hospitals on the English National Health Service website in 2010 using Weka data-mining software. We also compared the results obtained from sentiment analysis with the paper-based national inpatient survey results at the hospital level using Spearman rank correlation for all 161 acute adult hospital trusts in England.Results: There was 81%, 84%, and 89% agreement between quantitative ratings of care and those derived from free-text comments using sentiment analysis for cleanliness, being treated with dignity, and overall recommendation of hospital respectively (kappa scores:.40-.74, P<.001 for all). We observed mild to moderate associations between our machine learning predictions and responses to the large patient survey for the three categories examined (Spearman rho 0.37-0.51, P<.001 for all).Conclusions: The prediction accuracy that we have achieved using this machine learning process suggests that we are able to predict, from free-text, a reasonably accurate assessment of patients' opinion about different performance aspects of a hospital and that these machine learning predictions are associated with results of more conventional surveys.

Social media (for example Facebook and YouTube) uses online and mobile technologies to allow individuals to participate in, comment on and create user-generated content. Twitter is a widely used social media platform that lets users post short publicly available text-based messages called tweets that other users can respond to. Alongside traditional media outlets, Twitter has been a focus for discussions about the controversial and radical reforms to the National Health Service (NHS) in England that were recently passed into law by the current coalition Government. Looking at over 120,000 tweets made about the health reforms, we have investigated whether any insights can be obtained about the role of Twitter in informing, debating and influencing opinion in a specific area of health policy. In particular we have looked at how the sentiment of tweets changed with the passage of the Health and Social Care Bill through Parliament, and how this compared to conventional opinion polls taken over the same time period. We examine which users appeared to have the most influence in the ‘Twittersphere’ and suggest how a widely used metric of academic impact – the H-index – could be applied to measure context-dependent influence on Twitter.

BackgroundTraditional measures of patient experience have included surveys and, more recently, structured patient-reported outcome measures. There are also large amounts of unstructured, free-text information about the quality of health care available on the internet from blogs, social networks, and health-care rating websites that we are not scrutinising. In other industries, real-time natural language processing, such as sentiment analysis, of large datasets has provided a useful analytical approach to find patterns and understand data. If these techniques can be applied to health care, it opens up a novel approach to analyse large volumes of textual information about patient experience. The large number of free-text comments on the UK NHS Choices website allows an opportunity to examine these data through sentiment analysis. These comments are matched with the users' own quantitative ratings of the service, presenting an opportunity to measure the accuracy of natural language processing methods against the patient's own assessment. Simultaneously, the NHS has a developed programme of patient experience measurement via a national survey of hospital inpatients. Using these data sources, we have a natural opportunity to compare our sentiment analysis of comments to traditional patient surveys at an organisational level.MethodsWe tried to predict whether a patient would recommend a hospital, whether the hospital was clean, and whether they were treated with dignity from their free-text descriptions. We applied machine learning and natural language processing techniques to all (6400) online comments about hospitals on the NHS Choices website in 2010. We used open-source Weka data mining software. We used comments from NHS Choices data from 2008, 2009, and 2011 to train the software. Data from 2010 were used to test the predicting accuracy of the approach. We included our own a priori classification of the 1000 most common words and phrases in the analysis. Having calculated the accuracy of our prediction algorithm, we compared the results obtained with the national inpatient survey results for the same year (2010) at the hospital trust level with Spearman's test for rank correlation.FindingsWe were able to predict patients' rating of their care from their free-text comments with an accuracy of 81% for hospital cleanliness, 83% for treatment with dignity, and 89% for overall recommendation. We observed mild to moderate associations between our machine learning predictions and patient survey quantitative responses for the three categories examined: cleanliness (Spearman ρ=0·37, p<0·0001), dignity (ρ=0·50, p<0·0001), and overall recommendation (ρ=0·46, p<0·0001).InterpretationThe prediction accuracy that we have achieved using this machine learning process suggests that we are able to predict, from free text, a reasonably accurate assessment of patients' opinion about different performance aspects of a hospital. We also find that these machine learning predictions are associated to an extent with results of more conventional surveys. This work is ongoing and an iterative process, but suggests that it might be possible to monitor the so-called online cloud of patient experience in real-time and by doing so harness the value of patient opinion.FundingImperial College London is grateful for support from the National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care scheme, the National Institute for Health Research Biomedical Research Centre Funding scheme, and the Imperial Centre for Patient Safety and Service Quality.

Ad hoc networks can be formed from arbitrary collections of sensors, mobile routers, or business processes. These networks are open systems, in the sense that the network nodes share a common language but do not necessarily share a common goal or common knowledge, and there is no centralised controller or global data repository. Such systems have numerous advantages in terms of enabling autonomous, heterogeneous components to achieve individual goals without central direction and with only partial knowledge. However, operational problems stem from potential conflicts over resource allocation, miscommunication, and sub-ideal operation, and the general need of embedded systems to change behaviour according to changes in the environment. To address these problems, we propose to converge aspects of norm-governed specification from distributed multi-agent systems, opinion formation from social networks, and voting procedures from computational social choice. In particular, we develop a prototype system which interleaves gossiping, expressed preferences (voting) and norms, to configure rules and assign roles. This is another demonstration of the use of socially-inspired mechanisms for regulation of decentralised systems and a key step towards the realization of organized adaptation for open multi-agent systems.

Player classification has recently become a key aspect of game design in areas such as adaptive game systems, player behaviour prediction, player tutoring and non-player character design. Past research has focused on the design of hierarchical, preferencebased and probabilistic models aimed at modelling players' behaviour. We propose a meta-classification approach that breaks the clustering of gameplay mixed data into three levels of analysis. The first level uses dimensionality reduction and partitional clustering of aggregate game data in an action/skillbased classification. The second level applies similarity-based clustering of action sequences to group players according to their preferences. For this we propose a new approach which uses Rubner’s Earth Mover’s Distance (EMD) as a similarity metric to compare histograms of players’ game world explorations. The third level applies a combination of social network analysis metrics, such as shortest path length, to social data to find clusters in the players' social network. We test our approach in a gameplay dataset from a freely available first-person social hunting game.

Ad hoc networks can be formed from arbitrary collections of individual people (forming online computer-mediated communities), mobile routers (forming data communication networks) or electronic business processes (forming virtual enterprises). One way to deal with common features of dynamism in the network topology and membership, conflicts, sub-ideal operation, security, and the general need for continuous operation in the absence of a centralised facility, is to treat the ad hoc network as a normgoverned multi-agent system and use participatory adaptation as the mechanism for achieving autonomic capability (i.e. a global system response derived from the collective local behaviours and interactions of the individuals comprising the system). Therefore, complementing the formal representation of organisational behaviour defined in terms of roles, rules, norms, etc., this autonomic capability is at least partially derived from an underlying social network which plays a significant role in determining how, for example, conflicts are resolved and how the organisation itself is run. This position statement presents initial developments in what we call micro-social systems, which arise from interleaving a logical model of norm-governed systems with a mathematical model of social networks, and its application to issues of resource allocation, security, conflict resolution and self-adaptation in ad hoc networks.

Networked systems are the driving force of modern business and commerce, underpinned by ideas such as agile enterprises, holonic manufacturing, and dynamic real-time supply chains. On occasions, the system operation will be sub-optimal or non-ideal, and disputes will occur between independent partners. It may be undesirable to resolve such disputes by recourse to law; preferably, the parties in dispute would settle the matter by themselves. Therefore, we develop an alternative dispute resolution (ADR) system for virtual organizations as a way of settling disputes internally. We provide a norm-governed specification of an ADR protocol which is, effectively, an intelligent agent-based autonomic system. We develop this specification in two ways: concretely, through description of the mechanisms underlying protocol operation; and abstractly, by considering how the specification addresses principles for jury trials.

We present and analyse a model of opinion formation with dynamic confidence in agent-mediated social networks where the profiling of agents as leaders or followers is possible. An opinion leader is specified as a highly self- confident agent with strong opinions. An opinion follower is attracted to those agents in which it has more confidence. In our model, an agent i increases its confidence in another agent j based on how well j 's opinion meets the criteria specified in i's mind-set. A mind-set represents the set of beliefs, attitudes, assumptions and tendencies that predetermine the way an agent evaluates a received opinion. It is observed that the opinion formation in a group of persuadable agents, with similar confidence in each other, can easily lead to group think with agents following each other and one (or some) following the opinion of a single opinionated agent (i.e. an opinion leader). However this can be prevented by having at least one more self-confident, opinionated agent with an alternative opinion. This shows that divergent opinions from opinionated agents inhibit consensus. Furthermore, it is seen that once an equilibrium in the opinion formation has been reached, paradigm shifts can occur as the result of the sudden appearance of alternative opinion leaders.

Several different forms of peer-to-peer interactions, associations and interpersonal relations between human and artificial intelligences are described. We build upon a new form of grid computing which integrates human and artificial ‘processes’ in electronically saturated physical spaces, called socio-cognitive grids. We start from the analysis of three scenarios in P2P applications: digital rights management, mass user support and customer-to-customer interaction. These enable us to identify those factors that motivate the computing components in the socio-cognitive grids to form social structures, individually incorporating socio-cognitive intelligence and social awareness. In order to study the emergent properties of these social structures, such as reciprocity, social exchange and social networking, we need a theory that will help us understand the dynamics of social integration and support. We explore the use of a classical sociological theory of social structures and interpersonal relations. Subsequently we outline the components of a software simulation built on this theory and designed to formalize and evaluate this socio-computational intelligence. Ultimately our main aim is to analyse and understand those emergent properties that lead to the formation of stable and scalable social structures in socio-cognitive grids.