Krishnan Bhaskaran

Data Resource Profile: Clinical Practice Research Datalink (CPRD)

The Clinical Practice Research Datalink (CPRD) is an ongoing primary care database of anonymised medical records from general practitioners, with coverage of over 11.3 million patients from 674 practices in the UK. With 4.4 million active (alive, currently registered) patients meeting quality criteria, approximately 6.9% of the UK population are included and patients are broadly representative of the UK general population in terms of age, sex and ethnicity. General practitioners are the gatekeepers of primary care and specialist referrals in the UK. The CPRD primary care database is therefore a rich source of health data for research, including data on demographics, symptoms, tests, diagnoses, therapies, health-related behaviours and referrals to secondary care. For over half of patients, linkage with datasets from secondary care, disease-specific cohorts and mortality records enhance the range of data available for research. The CPRD is very widely used internationally for epidemiological research and has been used to produce over 1000 research studies, published in peer-reviewed journals across a broad range of health outcomes. However, researchers must be aware of the complexity of routinely collected electronic health records, including ways to manage variable completeness, misclassification and development of disease definitions for research.

Body-mass index and risk of 22 specific cancers: a population-based cohort study of 5·24 million UK adults

Summary Background High body-mass index (BMI) predisposes to several site-specific cancers, but a large-scale systematic and detailed characterisation of patterns of risk across all common cancers adjusted for potential confounders has not previously been undertaken. We aimed to investigate the links between BMI and the most common site-specific cancers. Methods With primary care data from individuals in the Clinical Practice Research Datalink with BMI data, we fitted Cox models to investigate associations between BMI and 22 of the most common cancers, adjusting for potential confounders. We fitted linear then non-linear (spline) models; investigated effect modification by sex, menopausal status, smoking, and age; and calculated population effects. Findings 5·24 million individuals were included; 166 955 developed cancers of interest. BMI was associated with 17 of 22 cancers, but effects varied substantially by site. Each 5 kg/m2 increase in BMI was roughly linearly associated with cancers of the uterus (hazard ratio [HR] 1·62, 99% CI 1·56–1·69; p<0·0001), gallbladder (1·31, 1·12–1·52; p<0·0001), kidney (1·25, 1·17–1·33; p<0·0001), cervix (1·10, 1·03–1·17; p=0·00035), thyroid (1·09, 1·00–1·19; p=0·0088), and leukaemia (1·09, 1·05–1·13; p≤0·0001). BMI was positively associated with liver (1·19, 1·12–1·27), colon (1·10, 1·07–1·13), ovarian (1·09, 1.04–1.14), and postmenopausal breast cancers (1·05, 1·03–1·07) overall (all p<0·0001), but these effects varied by underlying BMI or individual-level characteristics. We estimated inverse associations with prostate and premenopausal breast cancer risk, both overall (prostate 0·98, 0·95–1·00; premenopausal breast cancer 0·89, 0·86–0·92) and in never-smokers (prostate 0·96, 0·93–0·99; premenopausal breast cancer 0·89, 0·85–0·94). By contrast, for lung and oral cavity cancer, we observed no association in never smokers (lung 0·99, 0·93–1·05; oral cavity 1·07, 0·91–1·26): inverse associations overall were driven by current smokers and ex-smokers, probably because of residual confounding by smoking amount. Assuming causality, 41% of uterine and 10% or more of gallbladder, kidney, liver, and colon cancers could be attributable to excess weight. We estimated that a 1 kg/m2 population-wide increase in BMI would result in 3790 additional annual UK patients developing one of the ten cancers positively associated with BMI. Interpretation BMI is associated with cancer risk, with substantial population-level effects. The heterogeneity in the effects suggests that different mechanisms are associated with different cancer sites and different patient subgroups. Funding National Institute for Health Research, Wellcome Trust, and Medical Research Council.

Changes in the risk of death after HIV seroconversion compared with mortality in the general population

CONTEXT Mortality among human immunodeficiency virus (HIV)-infected individuals has decreased dramatically in countries with good access to treatment and may now be close to mortality in the general uninfected population. OBJECTIVE To evaluate changes in the mortality gap between HIV-infected individuals and the general uninfected population. DESIGN, SETTING, AND POPULATION Mortality following HIV seroconversion in a large multinational collaboration of HIV seroconverter cohorts (CASCADE) was compared with expected mortality, calculated by applying general population death rates matched on demographic factors. A Poisson-based model adjusted for duration of infection was constructed to assess changes over calendar time in the excess mortality among HIV-infected individuals. Data pooled in September 2007 were analyzed in March 2008, covering years at risk 1981-2006. MAIN OUTCOME MEASURE Excess mortality among HIV-infected individuals compared with that of the general uninfected population. RESULTS Of 16,534 individuals with median duration of follow-up of 6.3 years (range, 1 day to 23.8 years), 2571 died, compared with 235 deaths expected in an equivalent general population cohort. The excess mortality rate (per 1000 person-years) decreased from 40.8 (95% confidence interval [CI], 38.5-43.0; 1275.9 excess deaths in 31,302 person-years) before the introduction of highly active antiretroviral therapy (pre-1996) to 6.1 (95% CI, 4.8-7.4; 89.6 excess deaths in 14,703 person-years) in 2004-2006 (adjusted excess hazard ratio, 0.05 [95% CI, 0.03-0.09] for 2004-2006 vs pre-1996). By 2004-2006, no excess mortality was observed in the first 5 years following HIV seroconversion among those infected sexually, though a cumulative excess probability of death remained over the longer term (4.8% [95% CI, 2.5%-8.6%] in the first 10 years among those aged 15-24 years). CONCLUSIONS Mortality rates for HIV-infected persons have become much closer to general mortality rates since the introduction of highly active antiretroviral therapy. In industrialized countries, persons infected sexually with HIV now appear to experience mortality rates similar to those of the general population in the first 5 years following infection, though a mortality excess remains as duration of HIV infection lengthens.

Time series regression studies in environmental epidemiology

Time series regression studies have been widely used in environmental epidemiology, notably in investigating the short-term associations between exposures such as air pollution, weather variables or pollen, and health outcomes such as mortality, myocardial infarction or disease-specific hospital admissions. Typically, for both exposure and outcome, data are available at regular time intervals (e.g. daily pollution levels and daily mortality counts) and the aim is to explore short-term associations between them. In this article, we describe the general features of time series data, and we outline the analysis process, beginning with descriptive analysis, then focusing on issues in time series regression that differ from other regression methods: modelling short-term fluctuations in the presence of seasonal and long-term patterns, dealing with time varying confounding factors and modelling delayed (‘lagged’) associations between exposure and outcome. We finish with advice on model checking and sensitivity analysis, and some common extensions to the basic model.

Effects of ambient temperature on the incidence of myocardial infarction

Context: While the effects of weather and, in particular, ambient temperature on overall mortality are well documented, the strength of the evidence base for the effects on acute myocardial infarction (MI) are less clear. Objective: To systematically review studies specifically focusing on the effects of temperature on MI. Data sources: Medline, Embase, and GeoBase publication databases, as well as reference lists, and the websites of a number of relevant public organisations. Study selection: Studies of original data in which ambient temperature was an exposure of interest and MI a specific outcome were selected. Data extraction: The reported effects of ambient temperature on the risk of MI, including effect sizes and confidence intervals, where possible, were recorded. Methodological details were also extracted, including study population, location and setting, ascertainment of MI events, adjustment for potential confounders and consideration of lagged effects. Results: 19 studies were identified, of which 14 considered the short-term effects of temperature on a daily timescale, the remainder looking at longer-term effects. Overall, 8 of the 12 studies which included relevant data from the winter season reported a statistically significant short-term increased risk of MI at lower temperatures, while increases in risk at higher temperatures were reported in 7 of the 13 studies with relevant data. A number of differences were identified between studies in the population included demographics, location, local climate, study design and statistical methodology. Conclusion: A number of studies, including some that were large and relatively well controlled, suggested that both hot and cold weather had detrimental effects on the short-term risk of MI. However, further research with consistent methodology is needed to clarify the magnitude of these effects and to show which populations and individuals are vulnerable.

Changes in the incidence and predictors of human immunodeficiency virus-associated dementia in the era of highly active antiretroviral therapy.

Though effective anti–human immunodeficiency virus (HIV) therapies are now available, they have variable penetration into the brain. We therefore aimed to assess changes over calendar time in the risk for HIV‐associated dementia (HIV‐D), and factors associated with HIV‐D risk.

Short term effects of temperature on risk of myocardial infarction in England and Wales: time series regression analysis of the Myocardial Ischaemia National Audit Project (MINAP) registry

Objective To examine the short term relation between ambient temperature and risk of myocardial infarction. Design Daily time series regression analysis. Setting 15 conurbations in England and Wales. Participants 84 010 hospital admissions for myocardial infarction recorded in the Myocardial Ischaemia National Audit Project during 2003-6 (median 57 events a day). Main outcome measures Change in risk of myocardial infarction associated with a 1°C difference in temperature, including effects delayed by up to 28 days. Results Smoothed graphs revealed a broadly linear relation between temperature and myocardial infarction, which was well characterised by log-linear models without a temperature threshold: each 1°C reduction in daily mean temperature was associated with a 2.0% (95% confidence interval 1.1% to 2.9%) cumulative increase in risk of myocardial infarction over the current and following 28 days, the strongest effects being estimated at intermediate lags of 2-7 and 8-14 days: increase per 1°C reduction 0.6% (95% confidence interval 0.2% to 1.1%) and 0.7% (0.3% to 1.1%), respectively. Heat had no detrimental effect. Adults aged 75-84 and those with previous coronary heart disease seemed more vulnerable to the effects of cold than other age groups (P for interaction 0.001 or less in each case), whereas those taking aspirin were less vulnerable (P for interaction 0.007). Conclusions Increases in risk of myocardial infarction at colder ambient temperatures may be one driver of cold related increases in overall mortality, but an increased risk of myocardial infarction at higher temperatures was not detected. The risk of myocardial infarction in vulnerable people might be reduced by the provision of targeted advice or other interventions, triggered by forecasts of lower temperature.

Effects of air pollution on the incidence of myocardial infarction

Context: Short-term fluctuations in air pollution have been associated with changes in both overall and cardiovascular mortality. Objective: To consider the effects of air pollution on myocardial infarction (MI) risk by systematically reviewing studies looking at this specific outcome. Data sources: Medline, Embase and TOXNET publication databases, as well as reference lists and the websites of relevant public organisations. Study selection: Studies presenting original data with MI as a specific outcome and one or more of the following as an exposure of interest were included: particulate matter (PM), black carbon/black smoke, ozone, carbon monoxide, nitrogen oxides, sulphur dioxide and traffic exposure. Data extraction: The effects of each pollutant on risk of MI, including effect sizes and confidence intervals, were recorded where possible. Methodological details were also extracted including study population, location and setting, ascertainment of MI events, adjustment for potential confounders and consideration of lagged effects. Results: 26 studies were identified: 19 looked at the short-term effects of pollution on a daily timescale; the remaining 7 at longer-term effects. A proportion of studies reported statistically significant detrimental effects of PM with diameter <2.5 µm (3/5 studies, risk increase estimates ranging from 5 to 17% per 10 µg/m3 increase), PM <10 µm (3/10, 0.7–11% per 10 µg/m3), CO (6/14, 2–4% per ppm), SO2 (6/13, effect estimates on varied scales) and NO2 (6/13, 1–9% per 10 ppb). Increasing ozone levels were associated with a reduction in MI risk in 3/12 studies. A number of differences in location, population and demographics and study methodology between studies were identified that might have affected results. Conclusion: There is some evidence that short-term fluctuations in air pollution affect the risk of MI. However, further studies are needed to clarify the nature of these effects and identify vulnerable populations and individuals.

Completeness and usability of ethnicity data in UK-based primary care and hospital databases

Background Ethnicity recording across the National Health Service (NHS) has improved dramatically over the past decade. This study profiles the completeness, consistency and representativeness of routinely collected ethnicity data in both primary care and hospital settings. Methods Completeness and consistency of ethnicity recording was examined in the Clinical Practice Research Datalink (CPRD) and Hospital Episode Statistics (HES), and the ethnic breakdown of the CPRD was compared with that of the 2011 UK censuses. Results 27.1% of all patients in the CPRD (1990–2012) have ethnicity recorded. This proportion rises to 78.3% for patients registered since April 2006. The ethnic breakdown of the CPRD is comparable to the UK censuses. 79.4% of HES inpatients, 46.8% of outpatients and 26.8% of A&E patients had their ethnicity recorded. Amongst those with ethnicity recorded on >1 occasion, consistency was over 90% in all data sets except for HES inpatients. Combining CPRD and HES increased completeness to 97%, with 85% of patients having the same ethnicity recorded in both databases. Conclusions Using CPRD ethnicity from 2006 onwards maximizes completeness and comparability with the UK population. High concordance within and across NHS sources suggests these data are of high value when examining the continuum of care. Poor completeness and consistency of A&E and outpatient data render these sources unreliable.

The effects of hourly differences in air pollution on the risk of myocardial infarction: case crossover analysis of the MINAP database

Objectives To investigate associations between air pollution levels and myocardial infarction (MI) on short timescales, with data at an hourly temporal resolution. Design Time stratified case crossover study linking clinical data from the Myocardial Ischaemia National Audit Project (MINAP) with PM10, ozone, CO, NO2, and SO2 data from the UK National Air Quality Archive. Pollution effects were investigated with delays (lags) of 1–6, 7–12, 13–18, 19–24, and 25–72 hours in both single and multi-pollutant models, adjusted for ambient temperature, relative humidity, circulating levels of influenza and respiratory syncytial virus, day of week, holidays, and residual seasonality within calendar month strata. Setting Population based study in 15 conurbations in England and Wales. Subjects 79 288 diagnoses of myocardial infarction recorded over the period 2003–6. Main outcome measures Excess risk of myocardial infarction per 10 µg/m3 increase in pollutant level. Results In single pollutant models, PM10 and NO2 levels were associated with a very short term increase in risk of myocardial infarction 1–6 hours later (excess risks 1.2% (95% confidence interval 0.3 to 2.1) and 1.1% (0.3 to 1.8) respectively per 10 μg/m3 increase); the effects persisted in multi-pollutant models, though with only weak evidence of an independent PM10 effect (P=0.05). The immediate risk increases were followed by reductions in risk at longer lags: we found no evidence of any net excess risk associated with the five pollutants studied over a 72 hour period after exposure. Conclusions Higher levels of PM10 and NO2, which are typically markers of traffic related pollution, seem to be associated with transiently increased risk of myocardial infarction 1–6 hours after exposure, but later reductions in risk suggest that air pollution may be associated with bringing events forward in time (“short-term displacement”) rather than increasing overall risk. The well established effect of air pollution on cardiorespiratory mortality may not be mediated through increasing the acute risk of myocardial infarction, but through another mechanism.