The history of the Read Codes: the inaugural James Read Memorial Lecture 2011

Estimating the healthcare burden of Prurigo Nodularis in England: a CPRD database study

Purpose: Prurigo nodularis (PN) is a skin disease characterized by intensely itchy skin nodules and is associated with a significant healthcare resource utilization (HCRU). This study aimed to estimate the HCRU of patients in England with PN overall and moderate-to-severe PN (MSPN) in particular.

Methods: This retrospective cohort study used data from the Clinical Practice Research Datalink and Hospital Episode Statistics in England. Patients with Mild PN (MiPN) were matched to patients with MSPN by age and gender for the primary analysis. Patients were enrolled in the study between 1st April 2007 and 1st March 2019. All-cause HCRU was calculated, including primary and secondary care contacts and costs (cost-year 2022).

Results: Of 23,522 identified patients, 8,933 met the inclusion criteria, with a primary matched cohort of 2,479 PN patients. During follow up, the matched cohort's primary care visits were 21.27 per patient year (PPY) for MSPN group and 11.35 PPY for MiPN group. Any outpatient visits were 10.72 PPY and 4.87 PPY in MSPN and MiPN groups, respectively. Outpatient dermatology visits were 1.96 PPY and 1.14 PPY in MSPN and MiPN groups, respectively.

Conclusion: PN, especially MSPN, has a high HCRU burden in England, highlighting the need for new and improved disease management treatments.

The effect of herpes zoster vaccination at different stages of the disease course of dementia: Two quasi-randomized studies

The varicella zoster virus, a neurotropic herpesvirus, has been hypothesized to play a role in the pathophysiology of dementia, such as through neuroinflammatory processes or intracerebral vasculopathy. Using unique natural experiments, our group has previously found that live-attenuated herpes zoster (HZ) vaccination reduced the incidence of new diagnoses of dementia in both Wales and Australia. To inform further research and ultimately clinical care, it is crucial to understand at which stage of the disease course of dementia the HZ vaccine has its effect. Representing the two opposing ends of the dementia disease course as it can be ascertained from electronic health record data, the aims of this study were twofold: to determine the effect of HZ vaccination on i) new diagnoses of mild cognitive impairment (MCI) among individuals without any record of cognitive impairment, and ii) deaths due to dementia among individuals living with dementia. Our approach took advantage of the fact that at the time of the start date (September 1 2013) of the HZ vaccination program in Wales, individuals who had their eightieth birthday just after this date were eligible for HZ vaccination for one year whereas those who had their eightieth birthday just before were ineligible and remained ineligible for life. This eligibility rule created comparison groups just on either side of the September 2 1933 date-of-birth eligibility threshold who differed in their age by merely a week but had a large difference in their probability of receiving HZ vaccination. The key strength of our study is that these comparison groups should be similar in their health characteristics and behaviors except for a minute difference in age. We used regression discontinuity analysis to estimate the difference in our outcomes between individuals born just on either side of the date-of-birth eligibility threshold for HZ vaccination. Our dataset consisted of detailed country-wide electronic health record data from primary care in Wales, linked to hospital records and death certificates. We restricted our dataset to individuals born between September 1 1925 and September 1 1942. Among our study cohort of 282,557 without any record of cognitive impairment at baseline, HZ vaccination eligibility and receipt reduced the incidence of a new MCI diagnosis by 1.5 (95% CI: 0.5 - 2.9, p=0.006) and 3.1 (95% CI: 1.0 - 6.2, p=0.007) percentage points over nine years, respectively. Similarly, among our study cohort of 14,350 individuals who were living with dementia at baseline, being eligible for and receiving HZ vaccination reduced deaths due to dementia by 8.5 (95% CI: 0.6 - 18.5, p=0.036) and 29.5 (95% CI: 0.6 - 62.9, p=0.046) percentage points over nine years, respectively. Except for dementia, HZ vaccination did not have an effect on any of the ten most common causes of morbidity and mortality among adults aged 70 years and older in Wales in either of our two study cohorts. The protective effects of HZ vaccination for both MCI and deaths due to dementia were larger among women than men. Our findings suggest that the live-attenuated HZ vaccine has benefits for the dementia disease process at both ends of the disease course of dementia.

Assessing the uptake of incentivised physical health checks for people with serious mental illness: a cohort study in primary care

BACKGROUND

People with serious mental illness are more likely to experience physical illnesses. The onset of many of these illnesses can be prevented if detected early. Physical health screening for people with serious mental illness is incentivised in primary care in England through the Quality and Outcomes Framework (QOF). GPs are paid to conduct annual physical health checks on patients with serious mental illness, including checks of body mass index (BMI), cholesterol, and alcohol consumption.

AIM

To assess the impact of removing and reintroducing QOF financial incentives on uptake of three physical health checks (BMI, cholesterol, and alcohol consumption) for patients with serious mental illness.

DESIGN AND SETTING

Cohort study using UK primary care data from the Clinical Practice Research Datalink between April 2011 and March 2020.

METHOD

A difference-in-difference analysis was employed to compare differences in the uptake of physical health checks before and after the intervention, accounting for relevant observed and unobserved confounders.

RESULTS

An immediate change was found in uptake after physical health checks were removed from, and after they were added back to, the QOF list. For BMI, cholesterol, and alcohol checks, the overall impact of removal was a reduction in uptake of 14.3, 6.8, and 11.9 percentage points, respectively. The reintroduction of BMI screening in the QOF increased the uptake by 10.2 percentage points.

CONCLUSION

This analysis supports the hypothesis that QOF incentives lead to better uptake of physical health checks.

Familial hypercholesterolaemia in UK primary care: a Clinical Practice Research Datalink study of an under-recognised condition

BACKGROUND

Studies utilising genotyping methods report the prevalence of familial hypercholesterolaemia to be as high as one in 137 of the adult population.

AIM

To estimate the prevalence of familial hypercholesterolaemia measured by clinically coded diagnosis, associated treatments, and lipid measurements observed in UK primary care.

DESIGN AND SETTING

This was a retrospective analysis using the Clinical Practice Research Datalink (CPRD) GOLD database.

METHOD

Patients aged ≥18 years and actively registered on the index date (30 June 2018) formed the study cohort. Point prevalence of familial hypercholesterolaemia for 2018 was estimated overall and for each nation of the UK. Patients with familial hypercholesterolaemia were stratified into primary and secondary prevention groups, defined as those with/without a prior diagnosis of atherosclerotic cardiovascular disease. Prevalence estimates and extrapolations were replicated for these subgroups. Baseline demographic, lipid, and clinical characteristics for the prevalent cohort were presented.

RESULTS

In total, 4048 patients with familial hypercholesterolaemia formed the study cohort. The estimated familial hypercholesterolaemia prevalence for the UK was 16.4 per 10 000 (95% confidence interval [CI] = 16.0 to 16.9). Of these, 2646 (65.4%) patients with familial hypercholesterolaemia had a recent prescription for lipid-lowering therapy. Mean lipid levels were lower for those treated with lipid-lowering therapy compared with those untreated: 5.34 mmol/L (SD 1.50) versus 6.25 mmol/L (SD 1.55) for total cholesterol and 3.15 mmol/L (SD 1.34) versus 3.96 mmol/L (SD 1.36) for low-level density lipoprotein cholesterol.

CONCLUSION

The estimated prevalence of familial hypercholesterolaemia was one in 608 of the population, less than expected from other studies, which may indicate that familial hypercholesterolaemia is under-recognised in UK primary care. Over one-third of diagnosed patients were undertreated and many did not achieve target goals, placing them at risk of cardiovascular events.

Genetic influences on epilepsy outcomes: A whole-exome sequencing and health care records data linkage study

OBJECTIVE

This study was undertaken to develop a novel pathway linking genetic data with routinely collected data for people with epilepsy, and to analyze the influence of rare, deleterious genetic variants on epilepsy outcomes.

METHODS

We linked whole-exome sequencing (WES) data with routinely collected primary and secondary care data and natural language processing (NLP)-derived seizure frequency information for people with epilepsy within the Secure Anonymised Information Linkage Databank. The study participants were adults who had consented to participate in the Swansea Neurology Biobank, Wales, between 2016 and 2018. DNA sequencing was carried out as part of the Epi25 collaboration. For each individual, we calculated the total number and cumulative burden of rare and predicted deleterious genetic variants and the total of rare and deleterious variants in epilepsy and drug metabolism genes. We compared these measures with the following outcomes: (1) no unscheduled hospital admissions versus unscheduled admissions for epilepsy, (2) antiseizure medication (ASM) monotherapy versus polytherapy, and (3) at least 1 year of seizure freedom versus <1 year of seizure freedom.

RESULTS

We linked genetic data for 107 individuals with epilepsy (52% female) to electronic health records. Twenty-six percent had unscheduled hospital admissions, and 70% were prescribed ASM polytherapy. Seizure frequency information was linked for 100 individuals, and 10 were seizure-free. There was no significant difference between the outcome groups in terms of the exome-wide and gene-based burden of rare and deleterious genetic variants.

SIGNIFICANCE

We successfully uploaded, annotated, and linked genetic sequence data and NLP-derived seizure frequency data to anonymized health care records in this proof-of-concept study. We did not detect a genetic influence on real-world epilepsy outcomes, but our study was limited by a small sample size. Future studies will require larger (WES) data to establish genetic variant contribution to epilepsy outcomes.

Has COVID-19 affected dementia diagnosis rates in England?

BACKGROUND

The COVID-19 pandemic impacted on the provision of care and routine activity of all National Health Service (NHS) services. While General Practitioner referrals to memory services in England have returned to pre-pandemic levels, the estimated dementia diagnosis rate (DDR) fell by 5.4% between March 2020 and February 2023.

METHODS

In this paper we explore whether this reduction is accurate or is an artefact of the way the NHS collects data.

RESULTS

We explore the processes that may have affected national dementia diagnosis rates during and following the COVID-19 pandemic.

CONCLUSIONS

We discuss what action could be taken to improve the DDR in the future.

Causal evidence that herpes zoster vaccination prevents a proportion of dementia cases

The root causes of dementia are still largely unclear, and the medical community lacks highly effective preventive and therapeutic pharmaceutical agents for dementia despite large investments into their development. There is growing interest in the question if infectious agents play a role in the development of dementia, with herpesviruses attracting particular attention. To provide causal as opposed to merely correlational evidence on this question, we take advantage of the fact that in Wales eligibility for the herpes zoster vaccine (Zostavax) for shingles prevention was determined based on an individual's exact date of birth. Those born before September 2 1933 were ineligible and remained ineligible for life, while those born on or after September 2 1933 were eligible to receive the vaccine. By using country-wide data on all vaccinations received, primary and secondary care encounters, death certificates, and patients' date of birth in weeks, we first show that the percentage of adults who received the vaccine increased from 0.01% among patients who were merely one week too old to be eligible, to 47.2% among those who were just one week younger. Apart from this large difference in the probability of ever receiving the herpes zoster vaccine, there is no plausible reason why those born just one week prior to September 2 1933 should differ systematically from those born one week later. We demonstrate this empirically by showing that there were no systematic differences (e.g., in pre-existing conditions or uptake of other preventive interventions) between adults across the date-of-birth eligibility cutoff, and that there were no other interventions that used the exact same date-of-birth eligibility cutoff as was used for the herpes zoster vaccine program. This unique natural randomization, thus, allows for robust causal, rather than correlational, effect estimation. We first replicate the vaccine's known effect from clinical trials of reducing the occurrence of shingles. We then show that receiving the herpes zoster vaccine reduced the probability of a new dementia diagnosis over a follow-up period of seven years by 3.5 percentage points (95% CI: 0.6 - 7.1, p=0.019), corresponding to a 19.9% relative reduction in the occurrence of dementia. Besides preventing shingles and dementia, the herpes zoster vaccine had no effects on any other common causes of morbidity and mortality. In exploratory analyses, we find that the protective effects from the vaccine for dementia are far stronger among women than men. Randomized trials are needed to determine the optimal population groups and time interval for administration of the herpes zoster vaccine to prevent or delay dementia, as well as to quantify the magnitude of the causal effect when more precise measures of cognition are used. Our findings strongly suggest an important role of the varicella zoster virus in the etiology of dementia.

Clinical code usage in UK general practice: a cohort study exploring 18 conditions over 14 years

OBJECTIVE

To assess the diagnostic Read code usage for 18 conditions by examining their frequency and diversity in UK primary care between 2000 and 2013.

DESIGN

Population-based cohort study SETTING: 684 UK general practices contributing data to the Clinical Practice Research Datalink (CPRD) GOLD.

PARTICIPANTS

Patients with clinical codes for at least one of asthma, chronic obstructive pulmonary disease, diabetes, hypertension (HT), coronary heart disease, atrial fibrillation (AF), heart failure, stroke, hypothyroidism, chronic kidney disease, learning disability (LD), depression, dementia, epilepsy, severe mental illness (SMI), osteoarthritis, osteoporosis and cancer.

PRIMARY AND SECONDARY OUTCOME MEASURES

For the frequency ranking of clinical codes, canonical correlation analysis was applied to correlations of clinical code usage of 1, 3 and 5 years. Three measures of diversity (Shannon entropy index of diversity, richness and evenness) were used to quantify changes in incident and total clinical codes.

RESULTS

Overall, all examined conditions, except LD, showed positive monotonic correlation. HT, hypothyroidism, osteoarthritis and SMI codes' usage had high 5-year correlation. The codes' usage diversity remained stable overall throughout the study period. Cancer, diabetes and SMI had the highest richness (code lists need time to define) unlike AF, hypothyroidism and LD. SMI (high richness) and hypothyroidism (low richness) can last for 5 years, whereas cancer and diabetes (high richness) and LD (low richness) only last for 2 years.

CONCLUSIONS

This is an under-reported research area and the findings suggest the codes' usage diversity for most conditions remained overall stable throughout the study period. Generated mental health code lists can last for a long time unlike cardiometabolic conditions and cancer. Adopting more consistent and less diverse coding would help improve data quality in primary care. Future research is needed following the transfer to the Systematised Nomenclature of Medicine Clinical Terms (SNOMED CT) coding.

An Explainable Transformer-Based Deep Learning Model for the Prediction of Incident Heart Failure

Predicting the incidence of complex chronic conditions such as heart failure is challenging. Deep learning models applied to rich electronic health records may improve prediction but remain unexplainable hampering their wider use in medical practice. We aimed to develop a deep-learning framework for accurate and yet explainable prediction of 6-month incident heart failure (HF). Using 100,071 patients from longitudinal linked electronic health records across the U.K., we applied a novel Transformer-based risk model using all community and hospital diagnoses and medications contextualized within the age and calendar year for each patient's clinical encounter. Feature importance was investigated with an ablation analysis to compare model performance when alternatively removing features and by comparing the variability of temporal representations. A post-hoc perturbation technique was conducted to propagate the changes in the input to the outcome for feature contribution analyses. Our model achieved 0.93 area under the receiver operator curve and 0.69 area under the precision-recall curve on internal 5-fold cross validation and outperformed existing deep learning models. Ablation analysis indicated medication is important for predicting HF risk, calendar year is more important than chronological age, which was further reinforced by temporal variability analysis. Contribution analyses identified risk factors that are closely related to HF. Many of them were consistent with existing knowledge from clinical and epidemiological research but several new associations were revealed which had not been considered in expert-driven risk prediction models. In conclusion, the results highlight that our deep learning model, in addition high predictive performance, can inform data-driven risk factor identification.

Time trends in contraceptive prescribing in UK primary care 2000-2018: a repeated cross-sectional study

BACKGROUND

Over the last 20 years, new contraceptive methods became available and incentives to increase contraceptive uptake were introduced. We aimed to describe temporal trends in non-barrier contraceptive prescribing in UK primary care for the period 2000-2018.

METHODS

A repeated cross-sectional study using patient data from the IQVIA Medical Research Data (IMRD) database. The proportion (95% CI) of women prescribed non-barrier contraception per year was captured.

RESULTS

A total of 2 705 638 women aged 15-49 years were included. Between 2000 and 2018, the proportion of women prescribed combined hormonal contraception (CHC) fell from 26.2% (26.0%-26.3%) to 14.3% (14.2%-14.3%). Prescriptions for progestogen-only pills (POPs) and long-acting reversible contraception (LARC) rose from 4.3% (4.3%-4.4%) to 10.8% (10.7%-10.9%) and 4.2% (4.1%-4.2%) to 6.5% (6.5%-6.6%), respectively. Comparing 2018 data for most deprived versus least deprived areas, women from the most deprived areas were more likely to be prescribed LARC (7.7% (7.5%-7.9%) vs 5.6% (5.4%-5.8%)) while women from the least deprived areas were more likely to be prescribed contraceptive pills (20.8% (21.1%-21.5%) vs 26.2% (26.5%-26.9%)). In 2009, LARC prescriptions increased irrespective of age and social deprivation in line with a pay-for-performance incentive. However, following the incentive's withdrawal in 2014, LARC prescriptions for adolescents aged 15-19 years fell from 6.8% (6.6%-7.0%) in 2013 to 5.6% (5.4%-5.8%) in 2018.

CONCLUSIONS

CHC prescribing fell by 46% while POP prescribing more than doubled. The type of contraception prescribed was influenced by social deprivation. Withdrawal of a pay-for-performance incentive may have adversely affected adolescent LARC uptake, highlighting the need for further intervention to target this at-risk group.

Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers and risk of depression among older people with hypertension

BACKGROUND

Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs), commonly used antihypertensive drugs, may have a protective effect against depression in older individuals, but evidence in humans is limited.

AIMS

We evaluated the risk of depression, among older individuals with hypertension, comparing ACE or ARB initiators to thiazide(-like) diuretic initiators. Thiazide(-like) diuretics were used as control because these drugs are not associated with mood disorders.

METHODS

We used a propensity score-matched new user cohort design with routinely collected data from general practices in England from the Clinical Practice Research Datalink database. We matched 12,938 pairs of new users of ACEIs/ARBs and thiazide(-like) diuretics with hypertension (mean age 67.6 years; 54.7% women). Follow-up time started on the date of drug initiation and ended on the date of treatment discontinuation plus 30 days, or switch to a comparator, occurrence of a study event, death, date of patient's transfer out of practice, or end of the study period. The primary outcome was a composite endpoint of treated depression and nonfatal and fatal self-harm.

RESULTS/OUTCOMES

Compared to the thiazide(-like) diuretic group, ACEIs/ARBs use was not associated with a lower risk of the primary outcome (hazard ratio 0.96 (95% confidence interval: 0.79; 1.15)). Results did not differ according to lipophilicity, duration of use, and average daily dose, or class (ACEIs or ARBs).

CONCLUSIONS/INTERPRETATION

New use of ACEIs or ARBs is not associated with a lower risk of depression among individuals with hypertension.

Factors predicting amoxicillin prescribing in primary care among children: a cohort study

BACKGROUND

Antibiotic prescribing during childhood, most commonly for respiratory tract infections (RTIs), contributes to antimicrobial resistance, which is a major public health concern.

AIM

To identify factors associated with amoxicillin prescribing and RTI consultation attendance in young children in primary care.

DESIGN AND SETTING

Cohort study in Bradford spanning pregnancy to age 24 months, collected 2007-2013, linked to electronic primary care and air pollution data.

METHOD

Amoxicillin prescribing and RTI consultation rates/1000 child-years were calculated. Mixed-effects logistic regression models were fitted with general practice (GP) surgery as the random effect.

RESULTS

The amoxicillin prescribing rate among 2493 children was 710/1000 child-years during year 1 (95% confidence interval [CI] = 677 to 744) and 780/1000 (95% CI = 745 to 816) during year 2. During year 1, odds of amoxicillin prescribing were higher for boys (adjusted odds ratio [aOR] 1.36, 95% CI = 1.14 to 1.61), infants from socioeconomically deprived households (aOR 1.36, 95% CI = 1.00 to 1.86), and infants with a Pakistani ethnic background (with mothers born in the UK [aOR 1.44, 95% CI = 1.06 to 1.94] and outside [aOR 1.42, 95% CI = 1.07 to 1.90]). During year 2, odds of amoxicillin prescribing were higher for infants with a Pakistani ethnic background (with mothers born in the UK [aOR 1.46, 95% CI = 1.10 to 1.94] and outside [aOR 1.56, 95% CI = 1.19 to 2.04]) and those born <39 weeks gestation (aOR 1.20, 95% CI = 1.00 to 1.45). Additional risk factors included caesarean delivery, congenital anomalies, overcrowding, birth season, and childcare attendance, with GP surgery explaining 7%-9% of variation.

CONCLUSION

Socioeconomic status and ethnic background were associated with amoxicillin prescribing during childhood. Efforts to reduce RTI spread in household and childcare settings may reduce antibiotic prescribing in primary care.

UK poSt Arthroplasty Follow-up rEcommendations (UK SAFE): what does analysis of linked, routinely collected national data sets tell us about mid-late term revision risk after hip replacement? Retrospective cohort study

OBJECTIVE

To identify patients at risk of mid-late term revision of hip replacement to inform targeted follow-up.

DESIGN

Analysis of linked national data sets from primary and secondary care (Clinical Practice Research Datalink (CPRD-GOLD); National Joint Registry (NJR); English Hospital Episode Statistics (HES); Patient-Reported Outcome Measures (PROMs)).

PARTICIPANTS

Primary elective total hip replacement (THR) aged≥18.

EVENT OF INTEREST

Revision surgery≥5 years (mid-late term) after primary THR.

STATISTICAL METHODS

Cox regression modelling to ascertain risk factors of mid-late term revision. HR and 95% CI assessed association of sociodemographic factors, comorbidities, medication, surgical variables and PROMs with mid-late term revision.

RESULTS

NJR-HES-PROMs data were available from 2008 to 2011 on 142 275 THR; mean age 70.0 years and 61.9% female. CPRD GOLD-HES data covered 1995-2011 on 17 047 THR; mean age 68.4 years, 61.8% female. Patients had minimum 5 years postprimary surgery to end 2016. In NJR-HES-PROMS data, there were 3582 (2.5%) revisions, median time-to-revision after primary surgery 1.9 years (range 0.01-8.7), with 598 (0.4%) mid-late term revisions; in CPRD GOLD, 982 (5.8%) revisions, median time-to-revision 5.3 years (range 0-20), with 520 (3.1%) mid-late term revisions.Reduced risk of mid-late term revision was associated with older age at primary surgery (HR: 0.96; 95% CI: 0.95 to 0.96); better 6-month postoperative pain/function scores (HR: 0.35; 95% CI: 0.27 to 0.46); use of ceramic-on-ceramic (HR: 0.73; 95% CI: 0.56 to 0.95) or ceramic-on-polyethylene (HR: 0.76; 95% CI: 0.58 to 1.00) bearing surfaces.Increased risk of mid-late term revision was associated with the use of antidepressants (HR: 1.32; 95% CI: 1.09 to 1.59), glucocorticoid injections (HR: 1.33; 95% CI: 1.06 to 1.67) and femoral head size≥44 mm (HR: 2.56; 95% CI: 1.09 to 6.02)No association of gender, obesity or Index of Multiple Deprivation was observed.

CONCLUSION

The risk of mid-late term THR is associated with age at primary surgery, 6-month postoperative pain and function and implant factors. Further work is needed to explore the associations with prescription medications observed in our data.

Association between cardiometabolic disease multimorbidity and all-cause mortality in 2 million women and men registered in UK general practices

BACKGROUND

Myocardial infarction (MI), stroke and diabetes share underlying risk factors and commonalities in clinical management. We examined if their combined impact on mortality is proportional, amplified or less than the expected risk separately of each disease and whether the excess risk is explained by their associated comorbidities.

METHODS

Using large-scale electronic health records, we identified 2,007,731 eligible patients (51% women) and registered with general practices in the UK and extracted clinical information including diagnosis of myocardial infarction (MI), stroke, diabetes and 53 other long-term conditions before 2005 (study baseline). We used Cox regression to determine the risk of all-cause mortality with age as the underlying time variable and tested for excess risk due to interaction between cardiometabolic conditions.

RESULTS

At baseline, the mean age was 51 years, and 7% (N = 145,910) have had a cardiometabolic condition. After a 7-year mean follow-up, 146,994 died. The sex-adjusted hazard ratios (HR) (95% confidence interval [CI]) of all-cause mortality by baseline disease status, compared to those without cardiometabolic disease, were MI = 1.51 (1.49-1.52), diabetes = 1.52 (1.51-1.53), stroke = 1.84 (1.82-1.86), MI and diabetes = 2.14 (2.11-2.17), MI and stroke = 2.35 (2.30-2.39), diabetes and stroke = 2.53 (2.50-2.57) and all three = 3.22 (3.15-3.30). Adjusting for other concurrent comorbidities attenuated these estimates, including the risk associated with having all three conditions (HR = 1.81 [95% CI 1.74-1.89]). Excess risks due to interaction between cardiometabolic conditions, particularly when all three conditions were present, were not significantly greater than expected from the individual disease effects.

CONCLUSION

Myocardial infarction, stroke and diabetes were associated with excess mortality, without evidence of any amplification of risk in people with all three diseases. The presence of other comorbidities substantially contributed to the excess mortality risks associated with cardiometabolic disease multimorbidity.

Development and application of the ocular immune-mediated inflammatory diseases ontology enhanced with synonyms from online patient support forum conversation

BACKGROUND

Unstructured text created by patients represents a rich, but relatively inaccessible resource for advancing patient-centred care. This study aimed to develop an ontology for ocular immune-mediated inflammatory diseases (OcIMIDo), as a tool to facilitate data extraction and analysis, illustrating its application to online patient support forum data.

METHODS

We developed OcIMIDo using clinical guidelines, domain expertise, and cross-references to classes from other biomedical ontologies. We developed an approach to add patient-preferred synonyms text-mined from oliviasvision.org online forum, using statistical ranking. We validated the approach with split-sampling and comparison to manual extraction. Using OcIMIDo, we then explored the frequency of OcIMIDo classes and synonyms, and their potential association with natural language sentiment expressed in each online forum post.

FINDINGS

OcIMIDo (version 1.2) includes 661 classes, describing anatomy, clinical phenotype, disease activity status, complications, investigations, interventions and functional impacts. It contains 1661 relationships and axioms, 2851 annotations, including 1131 database cross-references, and 187 patient-preferred synonyms. To illustrate OcIMIDo's potential applications, we explored 9031 forum posts, revealing frequent mention of different clinical phenotypes, treatments, and complications. Language sentiment analysis of each post was generally positive (median 0.12, IQR 0.01-0.24). In multivariable logistic regression, the odds of a post expressing negative sentiment were significantly associated with first posts as compared to replies (OR 3.3, 95% CI 2.8 to 3.9, p < 0.001).

CONCLUSION

We report the development and validation of a new ontology for inflammatory eye diseases, which includes patient-preferred synonyms, and can be used to explore unstructured patient or physician-reported text data, with many potential applications.

Can decision support combat incompleteness and bias in routine primary care data?

OBJECTIVE

Routine primary care data may be used for the derivation of clinical prediction rules and risk scores. We sought to measure the impact of a decision support system (DSS) on data completeness and freedom from bias.

MATERIALS AND METHODS

We used the clinical documentation of 34 UK general practitioners who took part in a previous study evaluating the DSS. They consulted with 12 standardized patients. In addition to suggesting diagnoses, the DSS facilitates data coding. We compared the documentation from consultations with the electronic health record (EHR) (baseline consultations) vs consultations with the EHR-integrated DSS (supported consultations). We measured the proportion of EHR data items related to the physician's final diagnosis. We expected that in baseline consultations, physicians would document only or predominantly observations related to their diagnosis, while in supported consultations, they would also document other observations as a result of exploring more diagnoses and/or ease of coding.

RESULTS

Supported documentation contained significantly more codes (incidence rate ratio [IRR] = 5.76 [4.31, 7.70] P < .001) and less free text (IRR = 0.32 [0.27, 0.40] P < .001) than baseline documentation. As expected, the proportion of diagnosis-related data was significantly lower (b = -0.08 [-0.11, -0.05] P < .001) in the supported consultations, and this was the case for both codes and free text.

CONCLUSIONS

We provide evidence that data entry in the EHR is incomplete and reflects physicians' cognitive biases. This has serious implications for epidemiological research that uses routine data. A DSS that facilitates and motivates data entry during the consultation can improve routine documentation.

The Use of an Amino Acid Formula Containing Synbiotics in Infants with Cow's Milk Protein Allergy-Effect on Clinical Outcomes

Cow’s milk protein allergy (CMPA) is common and costly. Clinical trials of infants with CMPA have shown that the use of an amino acid formula containing pre- and probiotics (synbiotics) (AAF-Syn) may lead to significant reductions in infections, medication prescriptions and hospital admissions, compared to AAF without synbiotics. These effects have not yet been confirmed in real-world practice. This retrospective matched cohort study examined clinical and healthcare data from The Health Improvement Network database, from 148 infants with CMPA (54% male, mean age at diagnosis 4.69 months), prescribed either AAF-Syn (probiotic Bifidobacterium breve M16-V and prebiotics, including chicory-derived oligo-fructose and long-chain inulin) or AAF. AAF-Syn was associated with fewer symptoms (−37%, p < 0.001), infections (−35%, p < 0.001), medication prescriptions (−19%, p < 0.001) and healthcare contacts (−18%, p = 0.15) vs. AAF. Infants prescribed AAF-Syn had a significantly higher probability of achieving asymptomatic management without hypoallergenic formula (HAF) (adjusted HR 3.70, 95% CI 1.97–6.95, p < 0.001), with a shorter clinical course of symptoms (median time to asymptomatic management without HAF 1.35 years vs. 1.95 years). AAF-Syn was associated with potential cost-savings of £452.18 per infant over the clinical course of symptoms. These findings may be attributable to the effect of the specific synbiotic on the gut microbiome. Further research is warranted to explore this. This real-world study provides evidence consistent with clinical trials that AAF-Syn may produce clinical and healthcare benefits with potential economic impact.

Quantifying prescribed high dose opioids in the community and risk of overdose

Background

Opioid prescribing for a range of health issues is increasing globally. The risk of fatal and non-fatal overdose is increased among people prescribed strong opioids: in high doses in the context of polypharmacy (the use of multiple medications at the same time), especially with other sedatives; and among people with multiple morbidities including cardiorespiratory, hepatic and renal conditions. This study described and quantified the prescribing of strong opioids, comorbidities and other overdose risk factors among those prescribed strong opioids, and factors associated with high/very high opioid dosage in a regional health authority in Scotland as part of a wider service improvement exercise.

Methods

Participating practices ran searches to identify patients prescribed strong opioids and their characteristics, polypharmacy, and other overdose risk factors. Data were anonymised before being analysed at practice and patient-level. Morphine Equivalent Doses were calculated for patients based on drug/dose information and classed as Low/Medium/High/Very High. Descriptive statistics were generated on the strong opioid patient population and overdose risk factors. The relationship between the prescribing of strong opioids and practice/patient-level factors was investigated using linear and logistic regression models.

Results

Eighty-five percent (46/54) of GP practices participated. 12.4% (42,382/341,240) of individuals in participating practices were prescribed opioids and, of these, one third (14,079/42,382) were prescribed strong opioids. The most common comorbidities and overdose risk factors among strong opioid recipients were pain (67.2%), cardiovascular disease (43.2%), and mental health problems (39.3%). There was a positive significant relationship between level of social deprivation among practice caseload and level of strong opioid prescribing (p < 0.001). People prescribed strong opioids tended to be older (mean 59.7 years) and female (8638, 61.4%) and, among a subset of patients, age, gender and opioid drug class were significantly associated with prescribing of High/Very High doses.

Conclusions

Our findings have identified a large population at potential risk of prescription opioid overdose. There is a need to explore pragmatic models of tailored interventions which may reduce the risk of overdose within this group and clinical practice may need to be tightened to minimise overdose risk for individuals prescribed high dose opioids.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-11162-4.

Bisphosphonates to reduce bone fractures in stage 3B+ chronic kidney disease: a propensity score-matched cohort study

BACKGROUND

Bisphosphonates are contraindicated in patients with stage 4+ chronic kidney disease. However, they are widely used to prevent fragility fractures in stage 3 chronic kidney disease, despite a lack of good-quality data on their effects.

OBJECTIVES

The aims of each work package were as follows. Work package 1: to study the relationship between bisphosphonate use and chronic kidney disease progression. Work package 2: to study the association between using bisphosphonates and fracture risk. Work package 3: to determine the risks of hypocalcaemia, hypophosphataemia, acute kidney injury and upper gastrointestinal events associated with using bisphosphonates. Work package 4: to investigate the association between using bisphosphonates and changes in bone mineral density over time.

DESIGN

This was a new-user cohort study design with propensity score matching.

SETTING AND DATA SOURCES

Data were obtained from UK NHS primary care (Clinical Practice Research Datalink GOLD database) and linked hospital inpatient records (Hospital Episode Statistics) for work packages 1-3 and from the Danish Odense University Hospital Databases for work package 4.

PARTICIPANTS

Patients registered in the data sources who had at least one measurement of estimated glomerular filtration rate of < 45 ml/minute/1.73 m2 were eligible. A second estimated glomerular filtration rate value of < 45 ml/minute/1.73 m2 within 1 year after the first was requested for work packages 1 and 3. Patients with no Hospital Episode Statistics linkage were excluded from work packages 1-3. Patients with < 1 year of run-in data before index estimated glomerular filtration rate and previous users of anti-osteoporosis medications were excluded from work packages 1-4.

INTERVENTIONS/EXPOSURE

Bisphosphonate use, identified from primary care prescriptions (for work packages 1-3) or pharmacy dispensations (for work package 4), was the main exposure.

MAIN OUTCOME MEASURES

Work package 1: chronic kidney disease progression, defined as stage worsening or starting renal replacement. Work package 2: hip fracture. Work package 3: acute kidney injury, hypocalcaemia and hypophosphataemia identified from Hospital Episode Statistics, and gastrointestinal events identified from Clinical Practice Research Datalink or Hospital Episode Statistics. Work package 4: annualised femoral neck bone mineral density percentage change.

RESULTS

Bisphosphonate use was associated with an excess risk of chronic kidney disease progression (subdistribution hazard ratio 1.12, 95% confidence interval 1.02 to 1.24) in work package 1, but did not increase the probability of other safety outcomes in work package 3. The results from work package 2 suggested that bisphosphonate use increased fracture risk (hazard ratio 1.25, 95% confidence interval 1.13 to 1.39) for hip fractures, but sensitivity analyses suggested that this was related to unresolved confounding. Conversely, work package 4 suggested that bisphosphonates improved bone mineral density, with an average 2.65% (95% confidence interval 1.32% to 3.99%) greater gain in femoral neck bone mineral density per year in bisphosphonate users than in matched non-users.

LIMITATIONS

Confounding by indication was a concern for the clinical effectiveness (i.e. work package 2) data. Bias analyses suggested that these findings were due to inappropriate adjustment for pre-treatment risk. work packages 3 and 4 were based on small numbers of events and participants, respectively.

CONCLUSIONS

Bisphosphonates were associated with a 12% excess risk of chronic kidney disease progression in participants with stage 3B+ chronic kidney disease. No other safety concerns were identified. Bisphosphonate therapy increased bone mineral density, but the research team failed to demonstrate antifracture effectiveness.

FUTURE WORK

Randomised controlled trial data are needed to demonstrate antifracture efficacy in patients with stage 3B+ chronic kidney disease. More safety analyses are needed to characterise the renal toxicity of bisphosphonates in stage 3A chronic kidney disease, possibly using observational data.

STUDY REGISTRATION

This study is registered as EUPAS10029.

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. 25, No. 17. See the NIHR Journals Library website for further project information. The project was also supported by the National Institute for Health Research Biomedical Research Centre, Oxford.

Consultation analysis: use of free text versus coded text

General practice in the United Kingdom has been using electronic health records for over two decades, but coding clinical information remains poor. Lack of interest and training are considerable barriers preventing code use levels improvement. Tailored training could be the way forward, to break barriers in the uptake of coding; to do so it is paramount to understand coding use of the particular clinicians, to recognise their needs. It should be possible to easily assess text quantity and quality in medical consultations. A tool to measure these parameters, which could be used to tailor training needs and assess change, is demonstrated. The tool is presented and a preliminary study using a randomised sample of five recent consultations from thirteen different clinicians is used as an example. The tool, based on using a word processor and a spread-sheet, allowed quantitative analysis among clinicians while word clouds permitted a qualitative comparison between coded and free text. The average amount of free text per consultation was 68.2 words, (ranging from 25.4 and 130.2 among clinicians); an average of 6% of the text was coded (ranging from 0 to 13%). Patterns among clinicians could be identified. Using Word cloud, a different text use was demonstrated depending on its purpose. Some free text could be turned into code but nomenclature probably prevented some of the codings, like the expression of time. This proof of concept demonstrated that it is possible to calculate what percentage of consultations are coded and what codes are used. This allowed understanding clinicians’ preferences; training needs and gaps in nomenclature.

Exploring the health and service utilisation of general practice patients with a history of adverse childhood experiences (ACEs): an observational study using electronic health records

OBJECTIVES

To examine the relationships between adverse childhood experiences (ACEs), chronic health and health service utilisation among a sample of general practice patients.

DESIGN

Cross-sectional observational study using anonymised data from electronic health records for 763 patients.

SETTING

Four general practices in northwest England and North Wales.

OUTCOME MEASURES

Patient demographic data (age, gender); body mass index; self-reported smoking status; self-reported ACEs; diagnosis of chronic health conditions; current mental health problems; total number of service contacts and repeat medication use in the previous 6 months.

RESULTS

A history of ACEs (experiencing abuse or neglect as a child, and/or growing up in a household characterised by violence, substance use, mental health problems or criminal behaviour) was strongly independently associated with current mental health problems, smoking and chronic obstructive pulmonary disease, showing a dose-response relationship with level of ACE exposure. Medication use and contact were significantly greater among patients with high ACE exposure (≥4 ACEs), compared with those with no ACEs. However, contrary to findings from population studies, health service utilisation was not significantly different for patients with increased ACE exposure (1-3 ACEs) and their ACE-free counterparts.

CONCLUSIONS

Findings highlight the contribution ACEs make to unequal distributions of risk to health and well-being and patterns of health service use in the UK.

Are symptoms of insomnia in primary care associated with subsequent onset of dementia? A matched retrospective case-control study

Objective: There is evidence from neuroimaging studies of an association between insomnia and early dementia biomarkers, but observational studies have so far failed to show a clear association between insomnia and the later development of dementia. We investigated the association between dementia diagnosis and recording of insomnia symptoms 5-10 years earlier in primary care.Method: A case-control study using data from the Clinical Practice Research Datalink. 15,209 cases with dementia (either Alzheimer's, vascular, mixed or non-specific subtypes) at least 65 years old at time of diagnosis, were matched with the same number of controls on year of birth and gender. We ascertained the presence of insomnia symptoms during a five-year period starting 10 years before the index date. Odds ratios for developing dementia were estimated using logistic regression after controlling for hypnotic exposure and physical and mental health comorbidities.Results: The adjusted odds ratio for dementia in those with previous insomnia was 1.34 (95% CI = 1.20-1.50).Conclusion: There is an association between dementia and previous insomnia. It may be possible to incorporate insomnia into predictive tools for dementia.

Incidence and prevalence of inflammatory bowel disease in UK primary care: a population-based cohort study

OBJECTIVES

We describe temporal trends in the recorded incidence of inflammatory bowel disease (IBD) in UK primary care patients between 2000 and 2018.

DESIGN

A cohort study.

SETTING

The IQVIA Medical Research data (IMRD) primary care database.

PARTICIPANTS

All individuals registered with general practices contributing to IMRD during the period 01 January 2000-31 December 2018.

MAIN OUTCOME MEASURES

The primary outcome was the recorded diagnosis of IBD.

RESULTS

11 325 025 individuals were included and 65 700 IBD cases were identified, of which 22 560 were incident diagnoses made during the study period. Overall, there were 8077 incident cases of Crohn's disease (CD) and 12 369 incident cases of ulcerative colitis (UC). Crude incidence estimates of 'IBD overall', CD and UC were 28.6 (28.2 to 28.9), 10.2 (10.0 to 10.5) and 15.7 (15.4 to 15.9)/100 000 person years, respectively. No change in IBD incidence was observed for adults aged 17-40 years and children aged 0-9 years. However, for adults aged over 40 years, incidence fell from 37.8 (34.5 to 41.4) to 23.6 (21.3 to 26.0)/100 000 person years (average decrease 2.3% (1.9 to 2.7)/year (p<0.0001)). In adolescents aged 10-16 years, incidence rose from 13.1 (8.4 to 19.5) to 25.4 (19.5 to 32.4)/100 000 person years (average increase 3.0% (1.7 to 4.3)/year (p<0.0001)). Point prevalence estimates on 31 December 2018 for IBD overall, CD and UC were 725, 276 and 397 per 100 000 people, respectively.

CONCLUSIONS

This is one of the largest studies ever undertaken to investigate trends in IBD epidemiology. Although we observed stable or falling incidence of IBD in adults, our results are consistent with some of the highest reported global incidence and prevalence rates for IBD, with a 94% rise in incidence in adolescents. Further investigation is required to understand the aetiological drivers.

Twenty-year sociodemographic trends in lung cancer in non-smokers: A UK-based cohort study of 3.7 million people

INTRODUCTION

There are reports that lung cancer in non-smokers (LCINS) is increasing in the United Kingdom (UK) and other high-income countries but evidence from large-scale cohort studies to support this claim is limited.

MATERIAL AND METHODS

Using The Health Improvement Network (THIN) IQVIA™ Medical Research Data, we identified a cohort of 3,679,831 people from the UK self-reporting to their primary care physician as never or non-smokers. We estimated age-adjusted incidence rates for recorded lung cancer before (1998-2007) and after (2008-2018) the introduction of smoke-free legislation using multivariable Poisson regression. We also explored the impact of geographic location, social deprivation and urbanicity.

RESULTS

The analysis included 3,212 lung cancer events and 28 million person-years (PYs). Between 1998 and 2007, the age-adjusted rates in men declined by 9% per year (95 %CI: 7-11%) from an estimated 5.6 to 1.5 per 10,000 PYs and by 3% per year (95 %CI: 1-5%) between 2008 and 2018. These trends for men were similar across sociodemographic strata. Between 1998 and 2007, age-adjusted rates were stable for women at 1.5 per 10,000 PYs. However, there was evidence that time trends for women differed depending on levels of social deprivation with rates increasing by 5% per year (95 %CI: 2-9%) from an estimated 1.3-2.1 per 10,000 PYs for women living in the least socially deprived areas. Sex-specific time trends from 2008 to 2016 were broadly similar in a separate cohort of self-reported never smokers from UK Biobank with cancer events linked to national registries.

CONCLUSION

In summary, the incidence of LCINS has reduced or remained stable for most of the UK with the possible exception of women living in the least socially deprived areas.

GP coding behaviour for non-specific clinical presentations: a pilot study

Background

Clinical coding is an integral part of primary care. Disease incidence studies based on primary care electronic health records (EHRs) rely on the accuracy of these codes. Current code validation methods are not appropriate for non-specific conditions and provide limited information about GPs' decision-making behaviour around coding. Qualitative methods could offer insight into decision-making behaviour around coding of patients with non-specific conditions.

Aim

To investigate the decision-making behaviour of GPs when applying Read codes to non-specific clinical presentations, using Lyme disease as a case example.

Design & setting

A pilot study was undertaken, involving masked semi-structured interviews of eight GPs in the North West of England.

Method

Semi-structured interviews were carried out based on 11 clinical cases representative of Lyme disease presentations. Discrete answers were described descriptively. Interview transcripts were analysed using a thematic approach.

Results

Themes underpinning GPs’ coding behaviour included: GP personal and professional experience; clinical evidence; diagnostic uncertainty; professional integrity and defensive practice; and patient-sourced health information and beliefs. GPs placed Lyme disease on their differential diagnosis list for five cases; in only two cases would GPs select a Lyme disease related Read code.

Conclusion

GPs were reluctant to code with specific diagnostic Read codes when they were presented with patients with vague or unfamiliar symptomology. This masked questionnaire methodology offers a new approach to validate incidence figures, based on Read codes of non-specific conditions. The reluctance to code poses many problems for primary care EHRs research. Further research is needed to understand what drives GPs’ coding behaviour.

A descriptive epidemiological study of the incidence of newly diagnosed Lyme disease cases in a UK primary care cohort, 1998-2016

BACKGROUND

Primary care is likely to see the highest number of Lyme disease patients. Despite this, there is limited published data regarding Lyme disease patients accessing primary care in the UK. We aim to describe trends in the incidence of a new diagnosis, and demographics of patients identified in a primary care electronic health database.

METHODS

A descriptive epidemiological study of Lyme disease coded patients in UK primary care. 3725 patients coded for Lyme disease during 1998-2016 were identified within The Health Improvement Network (THIN). Incidence rates and the demographics of cases identified were described. Poisson regression was used to analyse socio-demographic characteristics of the cases.

RESULTS

There was an increase in annual crude incidence rates, peaking in 2015 at 5.47 (95% CI 4.85-6.14) cases per 100,000 population per year. Multivariable analysis showed there were significant differences in the ages of those affected, incidence of a new diagnosis rose as deprivation levels improved, and that there was a higher incidence of cases living in rural areas compared to urban areas. There was no significant difference between sexes for the UK. Cases were significantly more likely to identify with being white compared to the national population.

CONCLUSIONS

An increasing incidence of patients newly coded with Lyme disease related Read codes was identified using data from a UK national primary care database. By comparing these incidence figures with national laboratory-confirmed surveillance data, a multiplication factor of 2.35 (95%CI 1.81-2.88) can be calculated in order to estimate the annual number of cases seen in primary care. The significant socio-demographic variables associated with a Lyme disease diagnosis likely reflect a complex interplay of socio-economic issues, which needs to be further explored. Future work is needed to examine the treatment and management of patients within this database.

Investigating multisite pain as a predictor of self-reported falls and falls requiring health care use in an older population: A prospective cohort study

Older people are continuing to fall despite fall prevention guidelines targeting known falls’ risk factors. Multisite pain is a potential novel falls’ risk factor requiring further exploration. This study hypothesises that: (1) an increasing number of pain sites and widespread pain predicts self-reported falls and falls recorded in primary and secondary healthcare records; (2) those relationships are independent of known falls’ risk factors and putative confounders. This prospective cohort study linked data from self-completed questionnaires, primary care electronic health records, secondary care admission statistics and national mortality data. Between 2002–2005, self-completion questionnaires were mailed to community-dwelling individuals aged 50 years and older registered with one of eight general practices in North Staffordshire, UK(n = 26,129) yielding 18,497 respondents. 11,375 respondents entered the study; 4386 completed six year follow-up. Self-reported falls were extracted from three and six year follow-up questionnaires. Falls requiring healthcare were extracted from routinely collected primary and secondary healthcare data. Increasing number of pain sites increased odds of future 3 year (odds ratio 1.12 (95% confidence interval: 1.01–1.24)) and 6 year self-reported fall (odds ratio 1.02 (1.00–1.03)) and increased hazard of future fall requiring primary healthcare (hazard ratio 1.01 (1.00–1.03)). The presence of widespread pain increased odds of future 3 year (odds ratio 1.27 (0.92–1.75)) and 6 year fall (odds ratio 1.43(1.06–1.95)) and increased hazard of future fall requiring primary healthcare (hazard ratio 1.27(0.98–1.65)). Multisite pain was not associated with future fall requiring secondary care admission. Multisite pain must be included as a falls’ risk factor in guidelines to ensure clinicians identify their older patients at risk of falls and employ timely implementation of current falls prevention strategies.

Primary care risk stratification in COPD using routinely collected data: a secondary data analysis

Most clinical contacts with chronic obstructive pulmonary disease (COPD) patients take place in primary care, presenting opportunity for proactive clinical management. Electronic health records could be used to risk stratify diagnosed patients in this setting, but may be limited by poor data quality or completeness. We developed a risk stratification database algorithm using the DOSE index (Dyspnoea, Obstruction, Smoking and Exacerbation) with routinely collected primary care data, aiming to calculate up to three repeated risk scores per patient over five years, each separated by at least one year. Among 10,393 patients with diagnosed COPD, sufficient primary care data were present to calculate at least one risk score for 77.4%, and the maximum of three risk scores for 50.6%. Linked secondary care data revealed primary care under-recording of hospital exacerbations, which translated to a slight, non-significant cohort average risk score reduction, and an understated risk group allocation for less than 1% of patients. Algorithmic calculation of the DOSE index is possible using primary care data, and appears robust to the absence of linked secondary care data, if unavailable. The DOSE index appears a simple and practical means of incorporating risk stratification into the routine primary care of COPD patients, but further research is needed to evaluate its clinical utility in this setting. Although secondary analysis of routinely collected primary care data could benefit clinicians, patients and the health system, standardised data collection and improved data quality and completeness are also needed.

The effect of smoking on outcomes following primary total hip and knee arthroplasty: a population-based cohort study of 117,024 patients

Background and purpose - Smoking is a modifiable risk factor that may adversely affect postoperative outcomes. Healthcare providers are increasingly denying smokers access to total hip and knee arthroplasty (THA and TKA) until they stop smoking. Evidence supporting this is unclear. We assessed the effect of smoking on outcomes following arthroplasty.Patients and methods - We identified THAs and TKAs from the Clinical Practice Research Datalink, which were linked with datasets from Hospital Episode Statistics and the Office for National Statistics to identify outcomes. The effect of smoking on postoperative outcomes (complications, medications, revision, mortality, patient-reported outcome measures [PROMs]) was assessed using adjusted regression models.Results - We studied 60,812 THAs and 56,212 TKAs (11% smokers, 33% ex-smokers, 57% non-smokers). Following THA, smokers had an increased risk of lower respiratory tract infection (LRTI) and myocardial infarction compared with non-smokers and ex-smokers. Following TKA, smokers had an increased risk of LRTI compared with non-smokers. Compared with non-smokers (THA relative risk ratio [RRR] = 0.65; 95% CI = 0.61-0.69; TKA RRR = 0.82; CI = 0.78-0.86) and ex-smokers (THR RRR = 0.90; CI = 0.84-0.95), smokers had increased opioid usage 1-year postoperatively. Similar patterns were observed for weak opioids, paracetamol, and gabapentinoids. 1-year mortality rates were higher in smokers compared with non-smokers (THA hazard ratio [HR] = 0.37, CI = 0.29-0.49; TKA HR = 0.52, CI = 0.34-0.81) and ex-smokers (THA HR = 0.53, CI = 0.40-0.70). Long-term revision rates were not increased in smokers. Smokers had improvement in PROMs compared with preoperatively, with no clinically important difference in postoperative PROMs between smokers, non-smokers, and ex-smokers.Interpretation - Smoking is associated with more medical complications, higher analgesia usage, and increased mortality following arthroplasty. Most adverse outcomes were reduced in ex-smokers, therefore smoking cessation should be encouraged before arthroplasty.

Major adverse cardiovascular events in people with chronic kidney disease in relation to disease severity and diabetes status

Diabetes plays an important role in the complex relationship between chronic kidney disease (CKD) and cardiovascular disease. This retrospective observational study compared the influence of estimated glomerular filtration rate (eGFR) and proteinuria on the risk of major adverse cardiovascular event (MACE; myocardial infarction or stroke) in CKD patients with and without diabetes. Data were from a linked database of UK electronic health records. Individuals with CKD and no prior MACE were classified as type 1 diabetes (T1DM; n = 164), type 2 diabetes (T2DM; n = 9,711), and non-diabetes (non-DM; n = 75,789). Monthly updated time-dependent Cox proportional hazard models were constructed to calculate adjusted hazard ratios (aHRs) for progression to MACE from first record of abnormal eGFR or proteinuria (index date). In non-DM, aHRs (95% CIs) by baseline eGFR category (referent G2) were G1: 0.70 (0.55–0.90), G3a: 1.28 (1.20–1.35), G3b: 1.64 (1.52–1.76), G4: 2.19 (1.98–2.43), and G5: 3.12 (2.44–3.99), and by proteinuria category (referent A1) were A2: 1.13 (1.00–1.28), A2/3 (severity indeterminable): 1.58 (1.28–1.95), and A3: 1.64 (1.38–1.95). In T2DM, aHRs were G1: 0.98 (0.72–1.32), G3a: 1.18 (1.03–1.34), G3b: 1.31 (1.12–1.54), G4: 1.87 (1.53–2.29), G5: 2.87 (1.82–4.52), A2: 1.22 (1.04–1.42), A2/3: 1.45 (1.17–1.79), and A3: 1.82 (1.53–2.16). Low numbers in T1DM precluded analysis. Modelling T2DM and non-DM together, aHRs were, respectively, G1: 3.23 (2.38–4.40) and 0.70 (0.55–0.89); G2: 3.18 (2.73–3.70) and 1.00 (referent); G3a: 3.65 (3.13–4.25) and 1.28 (1.21–1.36); G3b: 4.01 (3.40–4.74) and 1.65 (1.54–1.77); G4: 5.78 (4.70–7.10) and 2.21 (2.00–2.45); G5: 9.00 (5.71–14.18) and 3.14 (2.46–4.00). In conclusion, reduced eGFR and proteinuria were independently associated with increased risk of MACE regardless of diabetes status. However, the risk of MACE in the same eGFR state was 4.6–2.4 times higher in T2DM than in non-DM.

Can patients with low health literacy be identified from routine primary care health records? A cross-sectional and prospective analysis

Background

People with low health literacy (HL) are at increased risk of poor health outcomes, and receive less benefit from healthcare services. However, healthcare practitioners can effectively adapt healthcare information if they are aware of their patients’ HL. Measurements are available to assess HL levels but may not be practical for use within primary care settings. New alternative methods based on demographic indicators have been successfully developed, and we aim to test if such methodology can be applied to routinely collected consultation records.

Methods

Secondary analysis was carried out from a recently completed prospective cohort study that investigated a primary care population who had consulted about a musculoskeletal pain problem. Participants completed questionnaires (assessing general health, HL, pain, and demographic information) at baseline and 6 months, with linked data from the participants’ consultation records. The Single Item Literacy Screener was used as a benchmark for HL. We tested the performance of an existing demographic assessment of HL, whether this could be refined/improved further (using questionnaire data), and then test the application in primary care consultation data. Tests included accuracy, sensitivity, specificity, and area under the curve (AUC). Finally, the completed model was tested prospectively using logistic regression producing odds ratios (OR) in the prediction of poor health outcomes (physical health and pain intensity).

Results

In total 1501 participants were included within the analysis and 16.1% were categorised as having low HL. Tests for the existing demographic assessment showed poor performance (AUC 0.52), refinement using additional components derived from the questionnaire improved the model (AUC 0.69), and the final model using data only from consultation data remained improved (AUC 0.64). Tests of this final consultation model in the prediction of outcomes showed those with low HL were 5 times more likely to report poor health (OR 5.1) and almost 4 times more likely to report higher pain intensity (OR 3.9).

Conclusions

This study has shown the feasibility of the assessment of HL using primary care consultation data, and that people indicated as having low HL have poorer health outcomes. Further refinement is now required to increase the accuracy of this method.

Coding algorithms for defining Charlson and Elixhauser co-morbidities in Read-coded databases

Background

Comorbidity measures, such as the Charlson Comorbidity Index (CCI) and Elixhauser Method (EM), are frequently used for risk-adjustment by healthcare researchers. This study sought to create CCI and EM lists of Read codes, which are standard terminology used in some large primary care databases. It also aimed to describe and compare the predictive properties of the CCI and EM amongst patients with hip fracture (and matched controls) in a large primary care administrative dataset.

Methods

Two researchers independently screened 111,929 individual Read codes to populate the 17 CCI and 31 EM comorbidity categories. Patients with hip fractures were identified (together with age- and sex-matched controls) from UK primary care practices participating in the Clinical Practice Research Datalink (CPRD). The predictive properties of both comorbidity measures were explored in hip fracture and control populations using logistic regression models fitted with 30- and 365-day mortality as the dependent variables together with tests of equality for Receiver Operating Characteristic (ROC) curves.

Results

There were 5832 CCI and 7156 EM comorbidity codes. The EM improved the ability of a logistic regression model (using age and sex as covariables) to predict 30-day mortality (AUROC 0.744 versus 0.686). The EM alone also outperformed the CCI (0.696 versus 0.601). Capturing comorbidities over a prolonged period only modestly improved the predictive value of either index: EM 1-year look-back 0.645 versus 5-year 0.676 versus complete record 0.695 and CCI 0.574 versus 0.591 versus 0.605.

Conclusions

The comorbidity code lists may be used by future researchers to calculate CCI and EM using records from Read coded databases. The EM is preferable to the CCI but only marginal gains should be expected from incorporating comorbidities over a period longer than 1 year.

Electronic supplementary material

The online version of this article (10.1186/s12874-019-0753-5) contains supplementary material, which is available to authorized users.

Identification of Incident Uterine Fibroids Using Electronic Medical Record Data

Introduction

Uterine fibroids are the most common benign tumors of the uterus and are associated with considerable morbidity. Diagnosis codes have been used to identify fibroid cases, but their accuracy, especially for incident cases, is uncertain.

Methods

We performed medical record review on a random sample of 617 women who received a fibroid diagnosis during 2012-2014 to assess diagnostic accuracy for incident fibroids. We developed 2 algorithms aimed at improving incident case-finding using classification and regression tree analysis that incorporated additional electronic health care data on demographics, symptoms, treatment, imaging, health care utilization, comorbidities and medication. Algorithm performance was assessed using medical record as gold standard.

Results

Medical record review confirmed 482 fibroid cases as incident, resulting a 78 percent positive predictive value (PPV) for incident cases based on diagnosis codes alone. Incorporating additional electronic data, the first algorithm classified 395 women with a pelvic ultrasound on diagnosis date but none before as incident cases. Of these, 344 were correctly classified, yielding an 87 percent PPV, 71 percent sensitivity, and 62 percent specificity. A second algorithm built on the first algorithm and further classified women based on a fibroid diagnosis code of 218.9 in 2 years after incident diagnosis and lower body mass index; yielded 93 percent PPV, 53 percent sensitivity, and 85 percent specificity.

Conclusions

Compared to diagnosis codes alone, our algorithms using fibroid diagnosis codes and additional electronic data improved identification of incident cases with higher PPV, and high sensitivity or specificity to meet different aims of future studies seeking to identify incident fibroids from electronic data.

LabRS: A Rosetta stone for retrospective standardization of clinical laboratory test results

Objective

Clinical laboratories in the United States do not have an explicit result standard to report the 7 billion laboratory tests results they produce each year. The absence of standardized test results creates inefficiencies and ambiguities for secondary data users. We developed and tested a tool to standardize the results of laboratory tests in a large, multicenter clinical data warehouse.

Methods

Laboratory records, each of which consisted of a laboratory result and a test identifier, from 27 diverse facilities were captured from 2000 through 2015. Each record underwent a standardization process to convert the original result into a format amenable to secondary data analysis. The standardization process included the correction of typos, normalization of categorical results, separation of inequalities from numbers, and conversion of numbers represented by words (eg, "million") to numerals. Quality control included expert review.

Results

We obtained 1.266 × 109 laboratory records and standardized 1.252 × 109 records (98.9%). Of the unique unstandardized records (78.887 × 103), most appeared <5 times (96%, eg, typos), did not have a test identifier (47%), or belonged to an esoteric test with <100 results (2%). Overall, these 3 reasons accounted for nearly all unstandardized results (98%).

Conclusion

Current results suggest that the tool is both scalable and generalizable among diverse clinical laboratories. Based on observed trends, the tool will require ongoing maintenance to stay current with new tests and result formats. Future work to develop and implement an explicit standard for test results would reduce the need to retrospectively standardize test results.

The internal validation of weight and weight change coding using weight measurement data within the UK primary care Electronic Health Record

PURPOSE

To use recorded weight values to internally validate weight status and weight change coding in the primary care Electronic Health Record (EHR).

PATIENTS AND METHODS

We included adult patients with weight-related Read codes recorded in the UK's Clinical Practice Research Datalink EHR between 2000 and 2017. Weight status codes were compared to weight values recorded on the same day and positive predictive values (PPVs) were calculated for commonly used codes. Weight change codes were validated using three methods: the percentage (%) difference in kilograms at the time of the code and 1) the previous weight measurement, 2) the weight predicted using linear regression, and 3) the historic mean weight. Weight change codes were validated if estimates were consistent across two out of three methods.

RESULTS

A total of 8,108,481 weight codes were recorded in 1,000,002 patients' EHR. Twice as many were recorded in females (n=5,208,593, 64%). The mean body mass index for "overweight" codes ranged from 31.9 kg/m2 to 46.9 kg/m2 and from 17.4 kg/m2 to 19.2 kg/m2 for "underweight" codes. PPVs for the most commonly used weight status codes ranged from 81.3% (80%-82.5%) to 99.3% (99.2%-99.4%). Across the estimation methods, and using only validated weight change codes, mean weight loss ranged from - 5.2% (SD 5.8%) to -7.9% (SD 7.3%) and mean weight gain from 4.2 % (SD 5.5%) to 7.9 % (SD 8.2%). The previous and predicted weight methods were most consistent.

CONCLUSION

We have developed an internationally applicable methodology to internally validate weight-related EHR coding by using available weight measurement data. We demonstrate the UK Read codes that can be confidently used to classify weight status and weight change in the absence of weight values. We provide the first evidence from primary care that a Read code for unexpected weight loss represents a mean loss of ≥ 5 % in a 6-month period, which was broadly consistent across age groups and gender.

Funding and policy incentives to encourage implementation of point-of-care C-reactive protein testing for lower respiratory tract infection in NHS primary care: a mixed-methods evaluation

OBJECTIVES

Utilisation of point-of-care C-reactive protein testing for lower respiratory tract infection has been limited in UK primary care, with costs and funding suggested as important barriers. We aimed to use existing National Health Service funding and policy mechanisms to alleviate these barriers and engage with clinicians and healthcare commissioners to encourage implementation.

DESIGN

A mixed-methods study design was adopted, including a qualitative survey to identify clinicians' and commissioners' perceived benefits, barriers and enablers post-implementation, and quantitative analysis of results from a real-world implementation study.

INTERVENTIONS

We developed a funding specification to underpin local reimbursement of general practices for test delivery based on an item of service payment. We also created training and administrative materials to facilitate implementation by reducing organisational burden. The implementation study provided intervention sites with a testing device and supplies, training and practical assistance.

RESULTS

Despite engagement with several groups, implementation and uptake of our funding specification were limited. Survey respondents confirmed costs and funding as important barriers in addition to physical and operational constraints and cited training and the value of a local champion as enablers.

CONCLUSIONS

Although survey respondents highlighted the clinical benefits, funding remains a barrier to implementation in UK primary care and appears not to be alleviated by the existing financial incentives available to commissioners. The potential to meet incentive targets using lower cost methods, a lack of policy consistency or competing financial pressures and commissioning programmes may be important determinants of local priorities. An implementation champion could help to catalyse support and overcome operational barriers at the local level, but widespread implementation is likely to require national policy change. Successful implementation may reproduce antibiotic prescribing reductions observed in research studies.

The anatomy of phenotype ontologies: principles, properties and applications

The past decade has seen an explosion in the collection of genotype data in domains as diverse as medicine, ecology, livestock and plant breeding. Along with this comes the challenge of dealing with the related phenotype data, which is not only large but also highly multidimensional. Computational analysis of phenotypes has therefore become critical for our ability to understand the biological meaning of genomic data in the biological sciences. At the heart of computational phenotype analysis are the phenotype ontologies. A large number of these ontologies have been developed across many domains, and we are now at a point where the knowledge captured in the structure of these ontologies can be used for the integration and analysis of large interrelated data sets. The Phenotype And Trait Ontology framework provides a method for formal definitions of phenotypes and associated data sets and has proved to be key to our ability to develop methods for the integration and analysis of phenotype data. Here, we describe the development and products of the ontological approach to phenotype capture, the formal content of phenotype ontologies and how their content can be used computationally.

Access to hospital and community palliative care for patients with advanced cancer: A longitudinal population analysis

BACKGROUND

The UK National Health Service is striving to improve access to palliative care for patients with advanced cancer however limited information exists on the level of palliative care support currently provided in the UK. We aimed to establish the duration and intensity of palliative care received by patients with advanced cancer and identify which cancer patients are missing out.

METHODS

Retrospective cancer registry, primary care and secondary care data were obtained and linked for 2474 patients who died of cancer between 2010 and 2012 within a large metropolitan UK city. Associations between the type, duration, and amount of palliative care by demographic characteristics, cancer type, and therapies received were assessed using Chi-squared, Mann-Whitney or Kruskal-Wallis tests. Multinomial multivariate logistic regression was used to assess the odds of receiving community and/or hospital palliative care compared to no palliative care by demographic characteristics, cancer type, and therapies received.

RESULTS

Overall 64.6% of patients received palliative care. The average palliative care input was two contacts over six weeks. Community palliative care was associated with more palliative care events (p<0.001) for a longer duration (p<0.001). Patients were less likely to receive palliative care if they were: male (p = 0.002), aged 80 years or over (p<0.05), diagnosed with lung cancer (p<0.05), had not received an opioid prescription (p<0.001), or had not received chemotherapy (p<0.001). Patients given radiotherapy were more likely to receive community only palliative care compared to no palliative care (Odds Ratio = 1.49, 95% Confidence Interval = 1.16-1.90).

CONCLUSION

Timely supportive care for cancer patients is advocated but these results suggest that older patients and those who do not receive anti-cancer treatment or opioid analgesics miss out. These patients should be targeted for assessment to identify unmet needs which could benefit from palliative care input.

Accuracy and utility of using administrative healthcare databases to identify people with epilepsy: a protocol for a systematic review and meta-analysis

INTRODUCTION

In an increasingly digital age for healthcare around the world, administrative data have become rich and accessible tools for potentially identifying and monitoring population trends in diseases including epilepsy. However, it remains unclear (1) how accurate administrative data are at identifying epilepsy within a population and (2) the optimal algorithms needed for administrative data to correctly identify people with epilepsy within a population. To address this knowledge gap, we will conduct a novel systematic review of all identified studies validating administrative healthcare data in epilepsy identification. We provide here a protocol that will outline the methods and analyses planned for the systematic review.

METHODS AND ANALYSIS

The systematic review described in this protocol will be conducted to follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. MEDLINE and Embase will be searched for studies validating administrative data in epilepsy published from 1975 to current (01 June 2018). Included studies will validate the International Classification of Disease (ICD), Ninth Revision (ICD-9) onwards (ie, ICD-9 code 345 and ICD-10 codes G40-G41) as well as other non-ICD disease classification systems used, such as Read Codes in the UK. The primary outcome will be providing pooled estimates of accuracy for identifying epilepsy within the administrative databases validated using sensitivity, specificity, positive and negative predictive values, and area under the receiver operating characteristic curves. Heterogeneity will be assessed using the I² statistic and descriptive analyses used where this is present. The secondary outcome will be the optimal administrative data algorithms for correctly identifying epilepsy. These will be identified using multivariable logistic regression models. 95% confidence intervals will be quoted throughout. We will make an assessment of risk of bias, quality of evidence, and completeness of reporting for included studies.

ETHICS AND DISSEMINATION

Ethical approval is not required as primary data will not be collected. Results will be disseminated in peer-reviewed journals, conference presentations and in press releases.

PROSPERO REGISTRATION

CRD42017081212.

Child and adolescent musculoskeletal pain (CAM-Pain) feasibility study: testing a method of identifying, recruiting and collecting data from children and adolescents who consult about a musculoskeletal condition in UK general practice

OBJECTIVES

Test a method of identifying, recruiting and collecting data from children and adolescents who consult their general practitioner about a musculoskeletal condition.

DESIGN

Prospective cohort feasibility study.

SETTING

13 general practices in West Midlands of England.

PARTICIPANTS

Patients aged 8-19 years who consult their general practice about a musculoskeletal condition. Patients were identified via a relevant musculoskeletal Read code entered at the point of consultation.

OUTCOME MEASURES

Feasibility was assessed in terms of study processes (recruitment rates), data collection procedures (duration, response variability), resource utilisation (mail-outs) and ethical considerations (acceptability).

RESULTS

From October 2016 to February 2017, an eligible musculoskeletal Read code was entered on 343 occasions, 202 patients were excluded (declined, n=153; screened not suitable, n=49) at the point of consultation. The remaining 141 patients were mailed an invitation to participate (41.1%); 46 patients responded to the invitation (response rate: 32.6%), of which 27 patients consented (consent rate: 19.1%). Participants mean age was 13.7 years (SD 2.7) and current pain intensity was 2.8 (SD 2.7). All participants completed the 6-week follow-up questionnaire. All participants found the interview questions to be acceptable and would consider participating in a similar study in the future. The majority of general practitioners/nurse practitioners, and all of the research nurses reported to be adequately informed about the study and found the study processes acceptable.

CONCLUSION

The expected number of participants were identified and invited, but consent rate was low (<20%) indicating that this method is not feasible (eg, for use in a large prospective study). Recruiting children and adolescents with musculoskeletal conditions in a primary care setting currently presents a challenge for researchers. Further work is needed to identify alternative ways to conduct studies in this population in order to address the current knowledge gap in this field.

Impact of the introduction of a universal childhood influenza vaccination programme on influenza-related admissions to paediatric intensive care units in England

Introduction

A universal childhood influenza vaccination programme was introduced in the UK in September 2013. We examine the impact of the gradual introduction of this programme on influenza-related paediatric intensive care unit (PICU) admission rates in England.

Methods

We extracted data on all influenza-related admissions to PICUs in England in resident children aged 0–15 years old between October 2003 and March 2017 from the Paediatric Intensive Care Audit Network (PICANet) database. We estimated influenza-associated PICU admission rates per 100 000 children by age group, sex and winter season (October to March), and used Poisson regression models to estimate incidence rate ratios (IRRs) in the winter seasons since the introduction of universal childhood vaccination compared with the two winters before the introduction of the programme (2011–2013).

Results

We identified 929 influenza-related PICU admissions among 873 children. 48.3% of admissions were among children aged less than 2 years old. The influenza-associated PICU admission rate was 1.32 per 100 000 children (95% CI 1.23 to 1.40). We identified a significant increase in influenza PICU admissions in the winters following the introduction of the universal childhood vaccination programme compared with the winters of 2010/2011–2012/2013 among children aged <5 years old: IRR 1.58 (1.05, 2.37) in children <1 year, 2.71 (1.43, 5.17) in 1 year-olds and 1.98 (1.18, 3.31) in children 2–4 years old. No significant difference was found among children aged 5–15 years.

Conclusion

The universal childhood influenza vaccination has not yet reduced the influenza-associated burden on PICUs in England during its early phase of introduction. Monitoring of influenza PICU admission rates needs to continue in England to assess the long-term impact of universal paediatric influenza vaccination. Linkage between PICANet and national infection surveillance databases would better enable such monitoring.

Data Resource: the Kent Integrated Dataset (KID)

Introduction

Electronic healthcare records from the UK are accessible to researchers via several platforms, but these platforms typically include data from a limited subset of health and care services. The Kent Integrated Dataset (KID) provides insight into system-wide health and care utilisation for the whole population of Kent and Medway.

Processes

The KID uses pseudonymisation-at-source to link patient-level records from services including general practices, hospitals, community health services and social care. Data is refreshed monthly and processes to monitor data quality have been developed.

Data contents

For each episode of care, the KID includes date of the episode, the type of service accessed, the cost of the episode and clinical information such as the health condition being treated and results of diagnostic tests. The dataset also includes contextual information such as the neighbourhood deprivation.

Conclusions

The KID is a unique and rich dataset available to researchers who are investigating a broad range of public health questions. It provides system-level insight into patient journeys and care utilisation and supports commissioning based on patient needs

Incidence and antibiotic prescribing for clinically diagnosed urinary tract infection in older adults in UK primary care, 2004-2014

Urinary tract infections (UTI) are an important cause of morbidity and antibiotic use in older adults but there are little data describing disease burden in primary care. The aim of this study was to estimate the incidence of clinically diagnosed UTI and examine associated empirical antibiotic prescribing. We conducted a retrospective observational study using linked health records from almost one million patients aged ≥65 years old, registered with 393 primary care practices in England. We estimated incidence of clinically diagnosed UTI between March 2004 and April 2014, and used multilevel logistic regression to examine trends in empiric antibiotic prescribing. Of 931,945 older adults, 196,358 (21%) had at least one clinically diagnosed UTI over the study period. In men, the incidence of clinically diagnosed UTI per 100 person-years at risk increased from 2.81 to 3.05 in those aged 65–74, 5.90 to 6.13 in those aged 75–84, and 8.08 to 10.54 in those aged 85+. In women, incidence increased from 9.03 to 10.96 in those aged 65–74, 11.35 to 14.34 in those aged 75–84, and 14.65 to 19.80 in those aged 85+. Prescribing of broad-spectrum antibiotics decreased over the study period. There were increases in the proportion of older men (from 45% to 74%) and women (from 55% to 82%) with UTI, prescribed a UTI specific antibiotic. There were also increases in the proportion of older men (42% to 69%) and women (15% to 26%) prescribed antibiotics for durations recommended by clinical guidelines. This is the first population-based study describing the burden of UTI in UK primary care. Our findings suggest a need to better understand reasons for increasing rates of clinically diagnosed UTI and consider how best to address this important clinical problem.

Using a linked database for epidemiology across the primary and secondary care divide: acute kidney injury

Background

NHS England has mandated the use in hospital laboratories of an automated early warning algorithm to create a consistent method for the detection of acute kidney injury (AKI). It generates an ‘alert’ based on changes in serum creatinine level to notify attending clinicians of a possible incident case of the condition, and to provide an assessment of its severity. We aimed to explore the feasibility of secondary data analysis to reproduce the algorithm outside of the hospital laboratory, and to describe the epidemiology of AKI across primary and secondary care within a region.

Methods

Using the Hampshire Health Record Analytical database, a patient-anonymised database linking primary care, secondary care and hospital laboratory data, we applied the algorithm to one year (1st January-31st December 2014) of retrospective longitudinal data. We developed database queries to modularise the collection of data from various sectors of the local health system, recreate the functions of the algorithm and undertake data cleaning.

Results

Of a regional population of 642,337 patients, 176,113 (27.4%) had two or more serum creatinine test results available, with testing more common amongst older age groups. We identified 5361 (or 0.8%) with incident AKI indicated by the algorithm, generating a total of 13,845 individual AKI alerts. A cross-sectional assessment of each patient’s first alert found that more than two-thirds of cases originated in the community, of which nearly half did not lead to a hospital admission.

Conclusion

It is possible to reproduce the algorithm using linked primary care, secondary care and hospital laboratory data, although data completeness, data quality and technical issues must be overcome. Linked data is essential to follow the significant proportion of people with AKI who transition from primary to secondary care, and can be used to assess clinical outcomes and the impact of interventions across the health system. This study emphasises that the development of data systems bridging across different sectors of the health and social care system can provide benefits for researchers, clinicians, healthcare providers and commissioners.

Electronic supplementary material

The online version of this article (doi:10.1186/s12911-017-0503-8) contains supplementary material, which is available to authorized users.

Consultation patterns of children and adolescents with knee pain in UK general practice: analysis of medical records

BACKGROUND

Knee problems are common in children and adolescents. Despite this, little is known about the epidemiology of knee problems in children and adolescents who consult in general practice. The aim of this study was to describe consultations by children and adolescents about knee problems in general practice, and examine patterns of patient presentations and consultations by age group, sex and area of socio-economic deprivation.

METHODS

Consultations records specific to the knee region were extracted from a general practice consultation database (CiPCA) over a one year period. Knee consultation codes were organised into 'symptom' or 'diagnosis' (sub-categorised: 'trauma', 'non-trauma') categories. Descriptive statistics were used to describe patient presentations and number of consultations overall, and stratified analysis carried out on age group, sex, and area of socio-economic deprivation.

RESULTS

Out of all musculoskeletal consultations, knee problems were the fourth most common patient presentation, responsible for the second highest number of consultations. Patient presentations and consultations increased up to age 12-15 years and then stabilised. Symptoms codes e.g. 'knee pain' were used more commonly than diagnosis codes e.g. 'knee sprain' overall. However, symptom code use declined as age increased, more symptom codes were used in girls compared to boys, and more diagnosis codes were used in patients from areas of high socio-economic deprivation.

CONCLUSIONS

This study provides insight into the epidemiology of knee problems in children and adolescents in general practice. Future research is needed to improve our understanding of the knee problems encountered by GPs, and the influence socio-economic deprivation has on consultations.

Affective concordance in couples: a cross-sectional analysis of depression and anxiety consultations within a population of 13,507 couples in primary care

BACKGROUND

Depression and anxiety are common and have a significant impact on the individual and wider society. One theory proposed to explain a heightened risk for depression and anxiety is affective concordance in couples (e.g. influence of shared mood states, shared health beliefs). Whilst research has shown concordance for severe psychiatric illnesses and general mood in couples, little attention has been given to concordance for common psychiatric conditions such as depression and anxiety. The aims of this study were to test affective concordance in couples and examine potential influences on concordance.

METHODS

Study design is a 1-year cross-sectional study of anxiety and depression consultations in primary care. Data were obtained from a validated primary care database of recorded consultations. Outcome was the presence of an anxiety or depression Read Code (GP recorded reason for consultation) in the female (within the couple dyad), and exposure was a recorded Read Code of anxiety or depression in the male. Logistic regression was used to test associations with odds ratios (OR) and 95% confidence intervals (95% CI) reported. Statistical adjustment was carried out on potential influences of concordance; age, environment (deprivation), healthcare behaviour (consultation frequency), and comorbidity.

RESULTS

A population of 13,507 couples were identified in which 927 people consulted for anxiety and 538 for depression. Logistic regression showed a 3 times increase in odds of an anxiety consultation in females if their male partner had also consulted OR 2.98 (95% CI 2.15 to 4.13). For depression females were over 4 times the odds of consulting if their male partner had also consulted OR 4.45 (95% CI 2.79 to 7.09). Adjustment within a multivariable model showed some reduction in odds; concordant anxiety was reduced to 2.5 times odds OR 2.48 (95%CI 1.76 to 3.50) and depression reduced to OR 3.39 (2.07 to 5.54).

CONCLUSION

Results show significant associations for affective concordance in couples. Factors influencing concordance are comorbidity and environmental factors, however reasons for deciding to consult (positive or negative) are unknown. This study highlights the patients' social context as a factor in consultations for anxiety and depression and gives support to the consideration of the patient's household as an influence on mental health.

Secondary Use of Data: Non-Interventional Study Best Practices in Planning and Protocol Development

Well established guidelines already exist that address best practices for Non-Interventional Study (NIS) design and methods. These guidelines provide advice on things to consider while designing a study and developing a protocol, but do not necessarily capture specific details related to the implementation of NIS. The intent of this paper is to propose a best practice for conducting secondary use of data NIS. We propose that the ideal implementation of a NIS should include the development of a strong Study Concept, followed by a detailed Protocol, Analysis Plan, Report, and considerations for Dissemination. We review and discuss common mistakes/pitfalls and key considerations at each step from concept to publication. In many cases in this review, we have also provided suggestions or accessible resources that researchers can apply as a "best practices" guide when planning, conducting, or reviewing this investigative method.

A Clinical Prediction Model to Assess Risk for Pancreatic Cancer Among Patients With New-Onset Diabetes

BACKGROUND & AIMS

Approximately 50% of all patients with pancreatic ductal adenocarcinoma (PDA) develop diabetes mellitus before their cancer diagnosis. Screening individuals with new-onset diabetes might allow earlier diagnosis of PDA. We sought to develop and validate a PDA risk prediction model to identify high-risk individuals among those with new-onset diabetes.

METHODS

We conducted a retrospective cohort study in a population representative database from the United Kingdom. Individuals with incident diabetes after the age of 35 years and 3 or more years of follow-up after diagnosis of diabetes were eligible for inclusion. Candidate predictors consisted of epidemiologic and clinical characteristics available at the time of diabetes diagnosis. Variables with P values <.25 in the univariable analyses were evaluated using backward stepwise approach. Model discrimination was assessed using receiver operating characteristic curve analysis. Calibration was evaluated using the Hosmer-Lemeshow test. Results were internally validated using a bootstrapping procedure.

RESULTS

We analyzed data from 109,385 patients with new-onset diabetes. Among them, 390 (0.4%) were diagnosed with PDA within 3 years. The final model (area under the curve, 0.82; 95% confidence interval, 0.75-0.89) included age, body mass index, change in body mass index, smoking, use of proton pump inhibitors, and anti-diabetic medications, as well as levels of hemoglobin A1C, cholesterol, hemoglobin, creatinine, and alkaline phosphatase. Bootstrapping validation showed negligible optimism. If the predicted risk threshold for definitive PDA screening was set at 1% over 3 years, only 6.19% of the new-onset diabetes population would undergo definitive screening, which would identify patients with PDA with 44.7% sensitivity, 94.0% specificity, and a positive predictive value of 2.6%.

CONCLUSIONS

We developed a risk model based on widely available clinical parameters to help identify patients with new-onset diabetes who might benefit from PDA screening.

The reporting and diagnosis of uterine fibroids in the UK: an observational study

Background

Uterine fibroids (UFs) are the most common benign tumour in women, and many undergo hysterectomy or uterus-preserving procedures (UPPs) to manage their symptoms. We aimed to validate the recording of UFs in a primary care database, The Health Improvement Network (THIN), and to determine the incidence of UFs in the UK.

Methods

In this observational study, women in THIN aged 15–54 years between January 2000 and December 2009 with no previous record of UFs, hysterectomy or UPPs were identified. Individuals were followed up until there was a Read code indicating UFs, they reached 55 years of age or died, or the study ended. Among those without a UF code, women were identified with a code for hysterectomy, UPPs or heavy menstrual bleeding (HMB). Anonymized patient profiles from each category were randomly selected and reviewed. Subsequently, primary care physicians were asked to complete questionnaires to verify the diagnosis for a randomly selected subgroup.

Results

In total, 737,638 women were identified who met the initial inclusion criteria. The numbers of women with a code for UFs, hysterectomy, UPPs and HMB were 9380, 11,002, 3220 and 60,915, respectively; the proportions of confirmed cases of UFs were 88.8, 29.7, 57.7 and 15.9 %. The estimated number of women with UFs was 23,140 (64.0 % without a recorded UF diagnosis). The overall incidence of UFs was 5.8 per 1000 woman-years.

Conclusions

UFs were confirmed in a high proportion of women with UF Read codes. However, almost two-thirds of cases were identified among women with a code for hysterectomy, UPPs or HMB. These results show that UFs are under-recorded in UK primary care, and suggest that primary care physicians tend to code the symptoms of UFs more often than the diagnosis.

Electronic supplementary material

The online version of this article (doi:10.1186/s12905-016-0320-8) contains supplementary material, which is available to authorized users.

Parkinson's disease and colorectal cancer risk-A nested case control study

BACKGROUND

A pro-inflammatory gut microbiota was described in both Parkinson's disease and colorectal cancer (CRC) and recently α-synuclein was demonstrated in the enteric nervous system. We sought to evaluate the association between Parkinson's disease and CRC.

METHODS

We conducted a nested case-control study using a large primary-care database. Cases were defined as all individuals with CRC. Up to 4 controls were matched with each case based on age, sex, practice-site and duration of follow-up. The primary exposure of interest was diagnosis of Parkinson's disease prior to CRC as well as disease duration, and Parkinson's specific therapies. The primary analysis was a conditional logistic-regression to estimate odds ratios (ORs) and 95% confidence interval (95%CI).

RESULTS

The study included 22,093 CRC cases and 85,833 matched controls. Past medical history of Parkinson's disease >1 year before index-date was associated with lower CRC risk (OR 0.74, 95%CI 0.59-0.94). The inverse association was more prominent among females compared to males (0.64, 95%CI 0.42-0.96 and 0.8, 95%CI 0.60-1.07, respectively). While patients who received no therapy or therapy with dopamine agonists had a non-significant decrease in cancer risk, patients who were treated with dopamine had a non-significant elevated cancer risk.

CONCLUSION

Parkinson's disease is inversely associated with CRC risk.