ObjectivesUnstructured free-text patient feedback contains rich information, and analysing these data manually would require a lot of personnel resources which are not available in most healthcare organisations.To undertake a systematic review of the literature on the use of natural language processing (NLP) and machine learning (ML) to process and analyse free-text patient experience data.MethodsDatabases were systematically searched to identify articles published between January 2000 and December 2019 examining NLP to analyse free-text patient feedback. Due to the heterogeneous nature of the studies, a narrative synthesis was deemed most appropriate. Data related to the study purpose, corpus, methodology, performance metrics and indicators of quality were recorded.ResultsNineteen articles were included. The majority (80%) of studies applied language analysis techniques on patient feedback from social media sites (unsolicited) followed by structured surveys (solicited). Supervised learning was frequently used (n=9), followed by unsupervised (n=6) and semisupervised (n=3). Comments extracted from social media were analysed using an unsupervised approach, and free-text comments held within structured surveys were analysed using a supervised approach. Reported performance metrics included the precision, recall and F-measure, with support vector machine and Naïve Bayes being the best performing ML classifiers.ConclusionNLP and ML have emerged as an important tool for processing unstructured free text. Both supervised and unsupervised approaches have their role depending on the data source. With the advancement of data analysis tools, these techniques may be useful to healthcare organisations to generate insight from the volumes of unstructured free-text data.

A best evidence topic was written according to a structured protocol. The question addressed was whether cryoanalgesia improves post-thoracotomy pain and recovery. Twelve articles were identified that provided the best evidence to answer the question. The authors, date, journal, study type, population, main outcome measures and results are tabulated. Reported measures were pain scores, additional opiate requirements, incidence of hypoesthesia and change in lung function. Half of the articles reviewed failed to demonstrate superiority of cryoanalgesia over other pain relief methods; however, additional opiate requirements were reduced in patients receiving cryoanalgesia. Change in lung function postoperatively was equivocal. Cryoanalgesia potentiated the incidence of postoperative neuropathic pain. Further analysis of the source of cryoanalgesia, duration, temperature obtained and extent of blockade revealed numerous discrepancies. Three studies utilized CO2 as the source of cryoanalgesia and four used nitrous oxide, but at differing temperatures and duration. Five studies did not reveal the source of cyroanalgesia. The number of intercostal nerves anaesthetized in each study varied. Seven articles anaesthetized three intercostal nerves, three articles used five intercostal nerves, one article used four intercostal nerves and one used one intercostal nerve at the thoracotomy site. Thoracotomy closure and site of area of chest drain insertion may have a role in postoperative pain; but only one article explained method of closure, and two articles mentioned placement of chest drain through blocked dermatomes. No causal inferences can be made by the above results as they are not directly comparable due to confounding variables between studies. Currently, the evidence does not support the use of cryoanalgesia alone as an effective method for relieving post-thoracotomy pain.

The diagnosis and management of an acute DVT is difficult and mistakes are often made. The cost to the National Health Service (NHS) of litigation arising from failure to diagnose and treat DVT early is substantial. Clinical diagnosis alone is often unreliable and a large proportion of DVT occurring in hospital are asymptomatic. In the United Kingdom, clinical scoring systems, D-dimer and ultrasound (US) imaging have all been adopted to aid diagnosis via DVT pathways. These pathways aim to exclude DVT only and often fail to actually address the cause of the symptoms once DVT is eventually cleared.

Patients undergoing surgery are at an increased risk of VTE. Since the early 1990s the prevention of VTE has been dominated by the administration of low-molecular weight heparin during admission. New oral anticoagulants have been extensively researched and have increased in popularity. This chapter reviews why surgical patients are at increased risk of VTE and summaries both the pharmacological and mechanical methods of prophylaxis available.

A best evidence topic was written according to a structured protocol. The question addressed was whether dissection of the pulmonary ligament during an upper lobectomy would result in improved outcomes. A total of 85 articles were found using the reported search, of which eight represented the best evidence to answer the clinical question. The authors, date, journal, study type, population, main outcome measures and results are tabulated. Reported measures were complications associated with dissection (atelectasis, bronchial stenosis, bronchial obstruction and bronchial deformation) and preservation (insufficient lung expansion, pooling of effusion and atelectasis) of the pulmonary ligament, ratio (%) of dead space in longitudinal axis (movement of nonoperated lobes), change in the angle (degrees) of main bronchus on the operated side, overall morbidity and mortality, overall survival and conversion rates. In a randomized control trial, the dissection of the pulmonary ligament revealed no significant difference in the dead space ratio or change in the angle of the main bronchus when compared with preservation. Dissection of the ligament, in theory, reduces the free space in the upper thorax by increasing the mobility of the residual lobes. Dissection of the ligament may lead to bronchial deformation, stenosis, obstruction or lobar torsion. Preservation of the ligament may prevent this complication by suppressing the upward movement of residual lobes. However, this may result in pleural effusion in the free thoracic space that may potentially become infected resulting in an empyema or bronchial fistula. Five large case series were analysed; three routinely dissected the pulmonary ligament and two did not. There was no observed difference in clinical outcomes between the two groups. There is no convincing evidence that dissection of the pulmonary ligament in an upper lobectomy significantly improves outcomes and reduces complications.