ObjectivesWe sought to use natural language processing to develop a suite of language models to capture key symptoms of severe mental illness (SMI) from clinical text, to facilitate the secondary use ...of mental healthcare data in research.DesignDevelopment and validation of information extraction applications for ascertaining symptoms of SMI in routine mental health records using the Clinical Record Interactive Search (CRIS) data resource; description of their distribution in a corpus of discharge summaries.SettingElectronic records from a large mental healthcare provider serving a geographic catchment of 1.2 million residents in four boroughs of south London, UK.ParticipantsThe distribution of derived symptoms was described in 23 128 discharge summaries from 7962 patients who had received an SMI diagnosis, and 13 496 discharge summaries from 7575 patients who had received a non-SMI diagnosis.Outcome measuresFifty SMI symptoms were identified by a team of psychiatrists for extraction based on salience and linguistic consistency in records, broadly categorised under positive, negative, disorganisation, manic and catatonic subgroups. Text models for each symptom were generated using the TextHunter tool and the CRIS database.ResultsWe extracted data for 46 symptoms with a median F1 score of 0.88. Four symptom models performed poorly and were excluded. From the corpus of discharge summaries, it was possible to extract symptomatology in 87% of patients with SMI and 60% of patients with non-SMI diagnosis.ConclusionsThis work demonstrates the possibility of automatically extracting a broad range of SMI symptoms from English text discharge summaries for patients with an SMI diagnosis. Descriptive data also indicated that most symptoms cut across diagnoses, rather than being restricted to particular groups.
ObjectivesTo identify negative symptoms in the clinical records of a large sample of patients with schizophrenia using natural language processing and assess their relationship with clinical ...outcomes.DesignObservational study using an anonymised electronic health record case register.SettingSouth London and Maudsley NHS Trust (SLaM), a large provider of inpatient and community mental healthcare in the UK.Participants7678 patients with schizophrenia receiving care during 2011.Main outcome measuresHospital admission, readmission and duration of admission.Results10 different negative symptoms were ascertained with precision statistics above 0.80. 41% of patients had 2 or more negative symptoms. Negative symptoms were associated with younger age, male gender and single marital status, and with increased likelihood of hospital admission (OR 1.24, 95% CI 1.10 to 1.39), longer duration of admission (β-coefficient 20.5 days, 7.6–33.5), and increased likelihood of readmission following discharge (OR 1.58, 1.28 to 1.95).ConclusionsNegative symptoms were common and associated with adverse clinical outcomes, consistent with evidence that these symptoms account for much of the disability associated with schizophrenia. Natural language processing provides a means of conducting research in large representative samples of patients, using data recorded during routine clinical practice.
Abstract Background Negative symptoms account for the greatest burden of illness among individuals with schizophrenia. These symptoms are an increasingly important target for therapy, especially ...since their presence predicts poor long-term clinical outcomes. However, development of practicable methods for their assessment has been difficult. In this study, we present a novel support vector machine learning method to see whether we could identify the presence of negative symptoms in electronic health records. Methods We used routinely collected clinical data from the Biomedical Research Council Case Register, South London and Maudsley NHS Trust. Data were obtained from the case records of 7678 adults with schizophrenia receiving care in 2011, of whom 1590 were inpatients. A training dataset of around 200 case records from this sample was analysed with the General Architecture for Text Engineering Machine Learning software package to develop a text-mining tool that was subsequently used to estimate the prevalence of negative symptoms in the whole sample. Multivariable logistic and multiple linear regression analyses were done to investigate the association of negative symptomatology with age, sex, relationship status, impairment of activities of daily living, and (for inpatients) length of hospital stay. Findings 4269 patients (55·7%) had at least one negative symptom documented. Negative symptoms were particularly associated with patients who were aged 20–29 years (all other age groups odds ratio OR and upper 95% CI limit <1·0, p<0·001), male (OR 1·29, 95% CI 1·17–1·44; p<0·001), and not in a relationship (1·31, 1·11–1·56; p=0·002). They were also associated with impairment of activities of daily living (1·35, 1·21–1·52; p<0·001) and increased likelihood of hospital admission (1·24, 1·10–1·39; p<0.001). Among inpatients, emotional withdrawal (β=30·0, 95% CI 15·6–44·4; p<0·001) and apathy (27·4, 1·8–53·1; p=0·036) were particularly associated with increased length of stay in hospital. Interpretation Using a machine learning approach, we were able to identify the presence of negative symptoms in electronic health records. The data suggest that negative symptoms are evident in most patients with schizophrenia and are associated with poor clinical outcomes. These findings highlight the need for the development of new treatments that can alleviate negative symptoms. Furthermore, the increasing use of electronic health records highlights an opportunity to adopt support vector machine learning text-mining approaches to obtain data for research and clinical decision support in other areas of medicine. Funding UK Medical Research Council, National Institute for Health Research, Roche.
Serious mental illness (SMI, including schizophrenia, schizoaffective disorder, and bipolar disorder) is associated with worse general health. However, admissions to general hospitals have received ...little investigation. We sought to delineate frequencies of and causes for non-psychiatric hospital admissions in SMI and compare with the general population in the same area.
Records of 18 380 individuals with SMI aged ⩾20 years in southeast London were linked to hospitalisation data. Age- and gender-standardised admission ratios (SARs) were calculated by primary discharge diagnoses in the 10th edition of the World Health Organization International Classification of Diseases (ICD-10) codes, referencing geographic catchment data.
Commonest discharge diagnosis categories in the SMI cohort were urinary conditions, digestive conditions, unclassified symptoms, neoplasms, and respiratory conditions. SARs were raised for most major categories, except neoplasms for a significantly lower risk. Hospitalisation risks were specifically higher for poisoning and external causes, injury, endocrine/metabolic conditions, haematological, neurological, dermatological, infectious and non-specific ('Z-code') causes. The five commonest specific ICD-10 diagnoses at discharge were 'chronic renal failure' (N18), a non-specific code (Z04), 'dental caries' (K02), 'other disorders of the urinary system' (N39), and 'pain in throat and chest' (R07), all of which were higher than expected (SARs ranging 1.57-6.66).
A range of reasons for non-psychiatric hospitalisation in SMI is apparent, with self-harm, self-neglect and/or reduced healthcare access, and medically unexplained symptoms as potential underlying explanations.
Background For many years, reflection has been considered good practice in medical education. In public health (PH), while no formal training or teaching of reflection takes place, it is expected as ...part of continuous professional development. This paper aims to identify reflective models useful for PH and to review published literature on the role of reflection in PH. The paper also aims to investigate the reported contribution, if any, of reflection by PH workers as part of their professional practice. Methods A review of the literature was carried out in order to identify reflective experience, either directly related to PH or in health education. Free text searches were conducted for English language papers on electronic bibliographic databases in September 2011. Thirteen papers met the inclusion criteria and were reviewed. Results There is limited but growing evidence to suggest reflection improves practice in disciplines allied to PH. No specific models are currently recommended or widely used in PH. Conclusions Health education literature has reflective models which could be applied to PH practice.
Background: Individuals with severe mental illness (SMI- schizophrenia, schizoaffective disorder and bipolar disorder) experience higher levels of morbidity and mortality than the general population. ...An important policy goal is to reduce this gap. Investigating the contributory role of physical illness is of key importance to unpacking the associations between SMI and detrimental outcomes, such as premature mortality and frequent hospital admissions. This thesis builds on recent advances to data extraction technologies to investigate the above. Objectives: • To describe the relative contributions of major disease groups to the gap in life expectancy between individuals with SMI and the general population. • To describe the most common reasons for admission to non-psychiatric hospitals by individuals with SMI and the relative frequencies of these admissions compared to the general population. • As a proof of principle for ascertaining meaningful symptom profiles from routine mental health record text fields, to describe the prospective association between number of recorded negative symptoms and mental healthcare outcomes (admission, duration of admission, and readmission) among individuals with schizophrenia. • To describe the association with mortality and hospitalisation for each of six symptom dimensions (positive, negative, manic, disorganisation, catatonic and depressive) extracted from the clinical records of individuals with SMI. Methods: Information for SMI cohorts were derived from the Clinical Record Interactive Search (CRIS), a de-identified electronic patient records data resource. Data on mortality were extracted using existing linkages between CRIS and death certification (Office for National Statistics). Life expectancy estimates were used to explain the contributions of specific causes of death to the gap. Using Hospital Episode Statistics data, frequencies of and causes for non-psychiatric hospital admissions in SMI were compared to those in the catchment general population. Symptoms within clinical record text fields were ascertained using a range of natural language processing algorithms, and were assessed for their associations with mortality and hospitalisation outcomes. Results: Natural causes accounted for 79.2% of lost life-years in women and 78.6% in men. Deaths from circulatory disorders accounted for more life-years lost in women than men (22.0% versus 17.4%, respectively), as did deaths from cancer (8.1% versus 0%), but the contribution from respiratory disorders was lower in women than men (13.7% versus 16.5%). Commonest discharge diagnosis categories were urinary conditions, digestive conditions, unclassified symptoms, neoplasms, and respiratory conditions. SARs were raised for most major categories, except neoplasms where risk was significantly lower. Hospitalisation risks were specifically higher for poisoning and external causes, injury, endocrine/metabolic conditions, haematological, neurological, dermatological, infectious and non- specific ('Z-code') causes. The five commonest specific ICD-10 diagnoses at discharge were 'chronic renal failure' (N18), a non-specific code (Z04), 'dental caries' (K02), 'other disorders of the urinary system' (N39), and 'pain in throat and chest' (R07), all of which were higher than expected (SARs ranging 1.57-6.66). Proof of concept analyses showed negative symptoms were associated with younger age, male gender and single marital status, and with increased likelihood of hospital admission (OR 1.24, 95% CI 1.10 to 1.39), longer duration of admission (β-coefficient 20.5 days, 95% CI 7.6-33.5), and increased likelihood of readmission following discharge (OR 1.58, 95% CI 1.28 to 1.95). Cox regression analyses detected significant effect of positive (HR 1.08, 95% CI 1.03- 1.16), negative (HR 1.09 95% CI 1.02- 1.16) and catatonic (HR 1.09 95% CI 1.03- 1.16) symptoms on mortality with adjustment of age, sex, employment, marital status and ethnicity. Linear regression analyses detected significant effect of manic (β-coefficient 0.09, 95% CI 0.02- 0.15), catatonic (β-coefficient 0.08, 95% CI 0.02- 0.15) and depressive (β-coefficient 0.14,95% CI 0.08- 0.21) symptoms on admission to non-mental health hospitals. Implications: Clinically, findings from this thesis confirm that SMI has a substantial negative impact on physical health, associated with increased mortality and morbidity. Of clinical relevance, the findings showed differences in burden depending on types of symptoms recorded in health records. From a policy context, the gap in life expectancy and increased non-psychiatric hospitalisation, accounted for by a broad range of causes, need to be addressed systematically. Interventions should focus on a whole system approach to improve health benefits for individuals of SMI.