Scientific
publications

BACKGROUND & AIMS: To examine the effect of malnutrition on clinical outcomes and healthcare resource use from initial diagnosis by a general practitioner (GP) in the UK. METHODS: 1000 records of malnourished patients were randomly selected from The Health Improvement Network database and matched with a sample of 996 patients' records with no previous history of malnutrition. Patients' outcomes and resource use were quantified for six months following diagnosis. RESULTS: Malnourished patients utilised significantly more healthcare resources (e.g. 18.90 versus 9.12 GP consultations; p < 0.001, and 13% versus 5% were hospitalised; p < 0.05). The six-monthly cost of managing the malnourished and non-malnourished group was £1753 and £750 per patient respectively, generating an incremental cost of care following a diagnosis of malnutrition of £1003 per patient. Thirteen percent and 2% of patients died in the malnourished and non-malnourished group respectively (p < 0.001). Independent predictors of mortality were: malnutrition (OR: 7.70); age (per 10 years) (OR: 10.46); and the Charlson Comorbidity Index Score (per unit score) (OR: 1.24). CONCLUSION: The healthcare cost of managing malnourished patients was more than twice that of managing non-malnourished patients, due to increased use of healthcare resources. After adjusting for age and comorbidity, malnutrition remained an independent predictor of mortality.

AIMS: In November 2005, the indications for nicotine replacement therapy (NRT) were broadened in the United Kingdom, making all forms available to 12-17-year-olds. This study aimed to estimate the effects of this change on prescribing of NRT to adolescents in England. DESIGN: Segmented regression analysis of primary care data on prescribing of nicotine replacement therapy using general additive mixed models. SETTING: A total of 350 general practices in England included in The Health Improvement Network (THIN), a database of UK electronic primary care records. PARTICIPANTS: All 12-17-year-old patients registered with a THIN practice in England. Measurements   Monthly rates of prescribing of NRT per 100 000 12-17-year-old patients registered with a THIN practice between January 2002 and June 2009. FINDINGS: Prescribing of NRT was highest in 16-17-year-olds and extremely low in 12-13-year-olds, generally fewer than 10 prescriptions per 100 000 adolescents per month. Girls tended to be prescribed NRT more frequently than boys. In 12-17-year-olds rates of prescribing were increasing prior to the broadening of indications, and there was no step change, but the rate of increase was lower, after the indications were broadened. CONCLUSIONS: Recent changes to the marketing licence for nicotine replacement therapy in England to include adolescents appear to have had little or no effect on prescribing practices.

PURPOSE: To quantify the possible role of confounding in our earlier estimates of the risk of falls amongst older people who have been prescribed antihypertensives. METHODS: Self-controlled case series analysis of 9862 individuals older than 60 years of age with a first fall recorded between years 2003 and 2006 and at least 12 months of recorded history before the first fall, taken from 386 UK general practices contributing data to The Health Improvement Network primary care database. Conditional Poisson regression to estimate incidence rate ratios for periods of exposure and non-exposure, adjusting for the effect of age. RESULTS: Incidence rate ratio of first fall in days 1-21 of the first episode of thiazide prescription is 2.80 (95%CI 1.7-4.57). For beta-blockers, the rate ratio in day 22 onwards is 1.23 (95%CI 1.02-1.48) in the very first episode of medication and 1.21 (95%CI 1.02-1.42) in subsequent episodes but is not significant in other periods of exposure. The incidence rate of first falls in day 22 of any episode of calcium channel blocker prescription is 0.75 (95%CI 0.60-0.92). CONCLUSIONS: The study provides evidence that at least some of our earlier estimates were subject to confounding. Nevertheless, thiazide prescription remains associated with an increased risk of first fall, and this effect is strongest in the first 3 weeks of prescription. Our study demonstrates that the case series method provides an effective way to assess the extent of residual confounding in case-control studies.

PURPOSE: A semi-automated high-dimensional propensity score (hd-PS) algorithm has been proposed to adjust for confounding in claims databases. The feasibility of using this algorithm in other types of healthcare databases is unknown. METHODS: We estimated the comparative safety of traditional non-steroidal anti-inflammatory drugs (NSAIDs) and selective COX-2 inhibitors regarding the risk of upper gastrointestinal bleeding (UGIB) in The Health Improvement Network, an electronic medical record (EMR) database in the UK. We compared the adjusted effect estimates when the confounders were identified using expert knowledge or the semi-automated hd-PS algorithm. RESULTS: Compared with the 411,616 traditional NSAID initiators, the crude odds ratio (OR) of UGIB was 1.50 (95%CI: 0.98, 2.28) for the 43,569 selective COX-2 inhibitor initiators. The OR dropped to 0.81 (0.52, 1.27) upon adjustment for known risk factors for UGIB that are typically available in both claims and EMR databases. The OR remained similar when further adjusting for covariates--smoking, alcohol consumption, and body mass index-that are not typically recorded in claims databases (OR 0.81; 0.51, 1.26) or adding 500 empirically identified covariates using the hd-PS algorithm (OR 0.78; 0.49, 1.22). Adjusting for age and sex plus 500 empirically identified covariates produced an OR of 0.87 (0.56, 1.34). CONCLUSIONS: The hd-PS algorithm can be implemented in pharmacoepidemiologic studies that use primary care EMR databases such as The Health Improvement Network. For the NSAID-UGIB association for which major confounders are well known, further adjustment for covariates selected by the algorithm had little impact on the effect estimate.

INTRODUCTION: The degree of generalisability of patient databases to the general population is important for interpreting database research. This report describes the representativeness of The Health Improvement Network (THIN), a UK primary care database, of the UK population. METHODS: Demographics, deprivation (Townsend), Quality and Outcomes Framework (QOF) condition prevalence and deaths from THIN were compared with national statistical and QOF 2006/2007 data. RESULTS: Demographics were similar although THIN contained fewer people aged under 25 years. Condition prevalence was comparable, e.g. 3.5% diabetes prevalence in THIN, 3.7% nationally. More THIN patients lived in the most affluent areas (23.5% in THIN, 20% nationally). Between 1990 and 2009, standardised mortality ratio ranged from 0.81 (95% CI: 0.39-1.49; 1990) to 0.93 (95% CI: 0.48-1.64; 1995). Adjusting for demographics/deprivation, the 2006 THIN death rate was 9.08/1000 population close to the national death rate of 9.4/1000 population. CONCLUSION: THIN is generalisable to the UK for demographics, major condition prevalence and death rates adjusted for demographics and deprivation.

BACKGROUND: Potential adverse effects of antidepressants during pregnancy have caused concern about their use. There are, however, very limited detailed data on patterns of antidepressant prescribing in pregnancy. OBJECTIVE: To examine secular trends in prescribing during pregnancy, to assess whether pregnancy is a major determinant for stopping antidepressants, and to identify characteristics of those who stopped antidepressants during pregnancy. METHOD: In this cohort study, we obtained data on 114,999 pregnant women (median age at delivery, 30.5 years [interquartile range, 26-34 years]) who had a live birth between 1992 and 2006 and 22,677 nonpregnant women from The Health Improvement Network primary care database, one of the largest sources of continuous anonymized primary care data in the United Kingdom and broadly representative of UK general practice. This database includes information on age, sex, medical diagnosis and symptoms, health promotion activities, referrals to secondary care, and prescriptions for each registered individual. The database also holds information about social deprivation as measured using quintiles of the Townsend score. We used Cox regression analysis to compare time to last prescription in pregnant versus nonpregnant women and to identify characteristics of those women who stopped antidepressants during pregnancy. RESULTS: Antidepressant prescribing in pregnancy increased nearly 4-fold from 1992 to 2006 (relative risk = 3.87; 95% CI, 1.73-8.66; P < .001). Since 2001, approximately 3% of the cohort received antidepressants at some stage during pregnancy. Selective serotonin reuptake inhibitors accounted for approximately 80% of the prescribed antidepressants. Antidepressants were more likely to be stopped in pregnant than in nonpregnant women, in particular during the first 6 weeks of pregnancy (hazard ratio = 5.19; 95% CI, 4.85-5.56; P < .001). Only 10% of women treated before pregnancy still received antidepressants at the start of the third trimester. In contrast, 35% of nonpregnant women were still treated after a similar time period. CONCLUSIONS: Although antidepressant prescribing in pregnancy increased nearly 4-fold from 1992 to 2006, pregnancy was a major determinant of cessation of antidepressant medication, and most women did not receive further antidepressant prescriptions beyond 6 weeks of gestation. This finding may be explained by concerns about potential adverse effects of the medications, even though these concerns need to be balanced against the potential harm of inadequate treatment of depression during pregnancy.

AIMS: To estimate the effect of the introduction of a new smoking cessation medication, varenicline, and the publication of guidance related to its use, on trends in prescribing of smoking cessation medications in England. DESIGN: Interrupted time series analysis of primary care data on prescribing of smoking cessation medication using autoregressive integrated moving average (ARIMA) modelling. SETTING: A total of 446 general practices included in The Health Improvement Network (THIN), a database of UK electronic primary care records. PARTICIPANTS: All primary care patients registered with a THIN practice in England. MEASUREMENTS: Monthly rates of prescribing of varenicline, nicotine replacement therapy (NRT) and bupropion per 100 000 patients registered with a THIN practice between June 2000 and June 2009. FINDINGS: NRT was the most commonly prescribed stop smoking medication, and bupropion the least frequently prescribed. After its introduction in December 2006 varenicline rapidly became the second most commonly prescribed drug. There was no statistically significant change in overall prescribing for smoking cessation medications after its introduction (P = 0.760), or after the publication of the related guidance in July 2007 (P = 0.134). CONCLUSIONS: Soon after being introduced in England, varenicline was widely prescribed; after nicotine replacement therapy it was the most commonly prescribed cessation medication. However, this does not appear to have increased overall rates of prescribing for smoking cessation medication.

AIM: To determine the rate of newly detected underlying disease in men receiving their first (index) phosphodiesterase type 5 inhibitor (PDE5i) prescription. METHODS: This non-interventional, retrospective study used anonymised patient records from UK general practices identified from the THIN database. Records of men aged ≥ 18 years, who received an index PDE5i prescription between January 1999 and June 2008 and with a continuous medical history (≥ 60 months) before the index prescription were included. Primary end-points were the prevalence of underlying disease prior to the index prescription and to establish the detection rate, defined as cumulative incidence of such a diagnosis in the 3 months following the index prescription. Assessments included comparison with age-matched controls, comparison with identical time periods immediately before and 1 year after, index prescription, and changes over time during the study period. Descriptive statistics, analysis of proportions and multivariate logistic regression analysis were used. RESULTS: Among the 24,708 patients receiving a PDE5i, the prevalence of any underlying diagnosis before the index prescription was 70.23%; prevalence of vasculogenic disease was highest (48.20%). The detection rate of any underlying disease was 11.53%, and again highest for vasculogenic disease (4.07%). Compared with an age-matched control population, the additional detection rate of an unknown underlying disease at PDE5i prescription was 45 for hypertension, 61 for hypercholesterolaemia, 38 for diabetes and 5 for hypogonadism per 10,000 men. CONCLUSION: Only a minority of men with erectile dysfunction have a previously undiagnosed important underlying disease that is uncovered at the time of an initial PDE5i prescription by a GP.

BACKGROUND: People with severe mental illnesses (SMI), including schizophrenia, are at increased risk of cardiovascular disease (CVD). Guidelines recommend regular CVD screening and in the United Kingdom, since 2004, General Practitioners are remunerated for annual reviews. OBJECTIVES: To compare annual rates of CVD screening provision in people with and without SMI between 2000 and 2008. METHOD: We identified 18,696 people with SMI and 95,512 people without SMI in the UK The Health Improvement Network (THIN) primary care database. We compared the rates of measurement of blood pressure (BP), glucose, cholesterol and body mass index (BMI). RESULTS: Prior to 2004, all people with SMI, were significantly less likely to receive each measurement, (including people above and below 60 years of age). In 2003; adjusted incidence rate ratios (95% CI) for screening in people with SMI under 60 years compared to people without SMI were: BMI: 0.62 (0.58-0.65); BP: 0.59 (0.56-0.62); glucose: 0.66 (0.61-0.70) and cholesterol: 0.54 (0.49-0.59). By 2007 people with SMI under 60 were equally likely receive a measurement of BMI: 1.00 (0.96-1.04), glucose: 1.00 (0.96-1.05) and cholesterol: 0.95 (0.90-1.0); the gap in screening for BP had narrowed 0.87 (0.83-0.90). However people with SMI over 60 years of age remained significantly less likely to be screened. There was little difference in screening according to social deprivation. CONCLUSIONS: In UK primary care, people with SMI over 60 years of age remain less likely than the general population to receive annual CVD screening despite higher risk of developing CVD.

OBJECTIVES: To evaluate the risk of myocardial infarction and death from coronary heart disease after discontinuation of low dose aspirin in primary care patients with a history of cardiovascular events. DESIGN: Nested case-control study. SETTING: The Health Improvement Network (THIN) database in the United Kingdom. PARTICIPANTS: Individuals aged 50-84 with a first prescription for aspirin (75-300 mg/day) for secondary prevention of cardiovascular outcomes in 2000-7 (n=39,513). MAIN OUTCOME MEASURES: Individuals were followed up for a mean of 3.2 years to identify cases of non-fatal myocardial infarction or death from coronary heart disease. A nested case-control analysis assessed the risk of these events in those who had stopped taking low dose aspirin compared with those who had continued treatment. RESULTS: There were 876 non-fatal myocardial infarctions and 346 deaths from coronary heart disease. Compared with current users, people who had recently stopped taking aspirin had a significantly increased risk of non-fatal myocardial infarction or death from coronary heart disease combined (rate ratio 1.43, 95% confidence interval 1.12 to 1.84) and non-fatal myocardial infarction alone (1.63, 1.23 to 2.14). There was no significant association between recently stopping low dose aspirin and the risk of death from coronary heart disease (1.07, 0.67 to 1.69). For every 1000 patients, over a period of one year there were about four more cases of non-fatal myocardial infarction among patients who discontinued treatment with low dose aspirin (recent discontinuers) compared with patients who continued treatment. CONCLUSIONS: Individuals with a history of cardiovascular events who stop taking low dose aspirin are at increased risk of non-fatal myocardial infarction compared with those who continue treatment.

BACKGROUND & AIMS: We investigated the effect of different prevention strategies against upper gastrointestinal bleeding (UGIB) in the general population and in patients on antithrombotic or anti-inflammatory treatments. METHODS: We performed a population-based, nested, case-control study using The Health Improvement Network UK primary care database. From 2000 to 2007, we identified 2049 cases of UGIB and 20,000 controls. The relative risk (RR) of UGIB associated with various gastroprotective agents was estimated by comparing current use (defined as use within 30 days of the index date) with nonuse in the previous year, using multivariate logistic regression. RESULTS: The adjusted RR of UGIB associated with current use of proton pump inhibitors (PPIs) for more than 1 month was 0.58 (95% confidence interval [CI], 0.42-0.79) among patients who received low-dose acetylsalicylic acid (ASA), 0.18 (95% CI, 0.04-0.79) for clopidogrel, 0.17 (95% CI, 0.04-0.76) for dual antiplatelet therapy, 0.48 (95% CI, 0.22-1.04) for warfarin, and 0.51 (95% CI, 0.34-0.78) for nonsteroidal anti-inflammatory drugs. The corresponding estimates for therapy with histamine-2-receptor antagonists (H2RAs) were more unstable, but tended to be of a smaller magnitude. In the general population, PPI use was associated with a reduced risk of UGIB compared with nonuse (RR, 0.80; 95% CI, 0.68-0.94); no such reduction was observed for H2RAs or nitrates. CONCLUSIONS: PPI use was associated with a lower risk of UGIB in the general population and in patients on antithrombotic or anti-inflammatory therapy compared with nonuse of PPIs. The reduction in risks of UGIB was smaller in H2RA than in PPI users.

BACKGROUND: Scabies is a contagious parasitic infestation which causes an allergic reaction to just a few mites. OBJECTIVES: To examine the epidemiology of scabies consultations in the U.K. by age, sex, region of the country and time. METHODS: Data were obtained from the THIN (The Health Information Network) database provided by EPIC U.K. in which each consultation episode is coded according to a 'READ' code. Data were available for 1997-2005 inclusive (9 years) for approximately 8·5% of the U.K. population from 12 regions of the U.K. We used harmonic analysis to model the data over time and derived cycle amplitudes and phases for each region in the U.K. RESULTS: There was a significantly greater infestation rate among females with a relative risk of 1·24 (P < 0·001) relative to males. The age group 10-19 years had the highest infestation rates (with rates of 4·55 per 1000 and 5·92 per 1000 for males and females, respectively). The middle age groups had the smallest infestation rates. A cycle with length of between 15 and 17 years was derived, and the phase times of the cycles for each region suggested a progression of the disease originating from the North East, spreading to Northern parts of the U.K. and then to the Midlands and the South. CONCLUSIONS: The study confirms earlier studies with regard to the age/sex distribution of scabies. We suggest a contagious pattern of spread of scabies infestation in the U.K. with an epidemic cycle length of 15-17 years.

BACKGROUND: Previous studies have shown that the incidence of idiopathic pulmonary fibrosis (IPF) is rising in the U.K. and U.S.A. Death registrations and primary care data were used to determine the current trends in IPF incidence in the U.K. Because routine clinical data sets were used, the term IPF clinical syndrome (IPF-CS) is used to describe individuals in this study. METHODS: Age- and stratum-specific death registration rates between 1968 and 2008 were calculated and these were applied to the 2008 population to generate annual standardised expected number of deaths. Annual mortality rate ratios were calculated using Poisson regression. Computerised primary care records were used to determine incidence rates of IPF-CS between 2000 and 2008 stratified by age, sex and geographical region, and survival rates between calendar periods were compared. RESULTS: Annual death certificate recording of IPF-CS rose sixfold across the study period from 0.92 per 100,000 in the 1968-1972 calendar periods to 5.10 per 100,000 in the 2006-2008 calendar period, and were higher in men and the older age groups. The incidence of IPF-CS in primary care increased by 35% from 2000 to 2008, with an overall incidence rate of 7.44 per 100,000 person-years (95% CI 7.12 to 7.77). Incidence was higher in men, the older population and in Northwest England. CONCLUSIONS: The incidence of IPF-CS in primary care and registered deaths from this cause in the U.K. continues to rise in the 21st century. The current findings suggest that there are >5000 new cases diagnosed each year in the U.K.

OBJECTIVE: To evaluate the performance of the QFractureScores for predicting the 10 year risk of osteoporotic and hip fractures in an independent UK cohort of patients from general practice records. DESIGN: Prospective cohort study. SETTING: 364 UK general practices contributing to The Health Improvement Network (THIN) database. PARTICIPANTS: 2.2 million adults registered with a general practice between 27 June 1994 and 30 June 2008, aged 30-85 (13 million person years), with 25,208 osteoporotic fractures and 12,188 hip fractures. MAIN OUTCOME MEASURES: First (incident) diagnosis of osteoporotic fracture (vertebra, distal radius, or hip) and incident hip fracture recorded in general practice records. RESULTS: Results from this independent and external validation of QFractureScores indicated good performance data for both osteoporotic and hip fracture end points. Discrimination and calibration statistics were comparable to those reported in the internal validation of QFractureScores. The hip fracture score had better performance data for both women and men. It explained 63% of the variation in women and 60% of the variation in men, with areas under the receiver operating characteristic curve of 0.89 and 0.86, respectively. The risk score for osteoporotic fracture explained 49% of the variation in women and 38% of the variation in men, with corresponding areas under the receiver operating characteristic curve of 0.82 and 0.74. QFractureScores were well calibrated, with predicted risks closely matching those across all 10ths of risk and for all age groups. CONCLUSION: QFractureScores are useful tools for predicting the 10 year risk of osteoporotic and hip fractures in patients in the United Kingdom.

PURPOSE: To evaluate whether exposure to Cytochrome P450, family 3, subfamily A, polypeptide 4 (CYP3A4)-inducing antiepileptics increases fracture risk compared to CYP3A4-non-inducing antiepileptics. METHODS: We performed a retrospective cohort study of initiators of antiepileptic agents using a UK medical record database (The Health Improvement Network) from 1995 to 2007. We considered an antiepileptic user an initiator if he or she had not received a prescription for an antiepileptic agent within the first year after entry in the database. Proportional hazards regression was used to calculate hazard ratios for fracture during long-term (≥ 6 months) exposure to CYP3A4 inducing versus CYP3A4 non-inducing antiepileptics. RESULTS: We identified 4077 initiators of CYP3A4-inducing antiepileptics and 6433 initiators of CYP3A4-non-inducing antiepileptics with at least 6 months of antiepileptic exposure. During 6006 person-years exposed to CYP3A4-inducing antiepileptics, 118 fractures were identified for an incidence rate of 1.96 (95% confidence interval (CI): 1.63-2.35) fractures per 100 person-years. During 7184 person-years exposed to CYP3A4-non-inducing antiepileptics, 127 fractures were identified, for an incidence rate of 1.77 (95% CI: 1.47-2.10) fractures per 100 person-years. The adjusted hazard ratio for CYP3A4-inducing antiepileptic versus CYP3A4-non-inducing antiepileptic was 1.21 (95% CI: 0.93-1.56). No duration-response relationship was evident. CONCLUSIONS: Our results do not support the hypothesis that CYP3A4 induction by antiepileptic agents increases the fracture risk. Further research will be needed to evaluate whether mechanisms other than CYP3A4 induction might explain some of the elevated risk of fractures associated with long-term use of antiepileptic agents.

OBJECTIVE: To investigate switching from thiazolidinediones, and predictors for switching treatment, after publication of a meta-analysis reporting an increased risk of myocardial infarction associated with rosiglitazone use. RESEARCH DESIGN AND METHODS: Using the Health Information Network (THIN) UK primary care database, the number of people with type 2 diabetes prescribed either thiazolidinedione, rosiglitazone (n = 10,062) or pioglitazone (n = 4454), and the rate of switching from thiazolidinediones (n = 3301 and 1106, respectively), were computed for each month, May 2006 to January 2008. The probability of switching post-publication, May 2007 to January 2008, was modelled by logistic regression in a forward stepwise model. Variables included demographics, history of ischaemic heart disease (IHD), heart failure (HF) or stroke, risk factors for IHD, glucose-lowering and cardiovascular drug use, HbA(1c) and diabetes duration. RESULTS: There was a sharp increase in switching from both thiazolidinediones in summer 2007; rosiglitazone prescription numbers then decreased while pioglitazone prescribing increased. Switching from rosiglitazone was associated with IHD [adjusted odds ratio (OR) 1.72; 95% confidence intervals (CI) 1.47-2.00], insulin treatment (OR 5.10; 95% CI 3.21-8.10), HF (OR 2.26; 95% CI 1.62-3.18), a recent sulphonylurea prescription (OR 1.33; 95% CI 1.17-1.51) gender (OR men vs. women 0.79; 95% CI 0.70, 0.90) and duration of therapy. Switching from pioglitazone was associated with HF (OR 3.05; 95% CI 1.77-5.26), duration of therapy, and number of glucose-lowering treatments. CONCLUSIONS: Prescribing habits for both thiazolidinediones changed immediately following the safety warning. IHD was associated with switching from rosiglitazone; otherwise reasons for change appear to be complex, not directly related to the findings of the meta-analysis.