Catherine Welch

Assessment of an incentivised scheme to provide annual health checks in primary care for adults with intellectual disability: a longitudinal cohort study

BACKGROUND People with intellectual disabilities (ID) have many comorbidities but experience inequities in access to health care. National Health Service England uses an opt-in incentive scheme to encourage annual health checks of patients with ID in primary care. We investigated whether the first 3 years of the programme had improved health care of people with ID. METHODS We did a longitudinal cohort study that used data from The Health Improvement Network primary care database. We did multivariate logistic regression to assess associations between various characteristics and whether or not practices had opted in to the incentivised scheme. FINDINGS We assessed data for 8692 patients from 222 incentivised practices and those for 918 patients in 48 non-incentivised practices. More blood tests (eg, total cholesterol, odds ratio [OR] 1·88, 95% CI 1·47-2·41, p<0·0001) general health measurements (eg, smoking status, 6·0, 4·10-8·79, p<0·0001), specific health assessments (eg, hearing, 24·0, 11·5-49·9, p<0·0001), and medication reviews (2·23, 1·68-2·97, p<0·0001) were done in incentivised than in non-incentivised practices, and more health action plans (6·15, 1·41-26·9, p=0·0156) and secondary care referrals (1·47, 1·05-2·05, p=0·0256) were made. Identification rates were higher in incentivised practices for thyroid disorder (OR 2·72, 95% CI 1·09-6·81, p=0·0323), gastrointestinal disorders (1·94, 1·03-3·65, p=0·0390), and obesity (2·49, 1·76-3·53, p<0·0001). INTERPRETATION Targeted annual health checks for people with ID in primary care could reduce health inequities. FUNDING National Institute for Health Research.

Outcomes in UK patients with hospital-acquired bacteraemia and the risk of catheter-associated urinary tract infections.

Purpose There is lack of contemporary outcome data on patients with hospital-acquired infections that cause bacteraemia. We determined the risk factors for 7-day mortality and investigated the hypothesis that, compared with central venous catheter (CVC)-associated bacteraemic infections, catheter-associated bacteraemic urinary tract infections (UTIs) were significantly associated with 7-day mortality. Methods From October 2007 to September 2008, demographical, clinical and microbiological data were collected on patients with hospital-acquired bacteraemia. Patients were followed until death, hospital discharge or recovery from infection. Risk factors for 7-day mortality were determined and multivariate logistic regression was used to define the association between catheter-associated bacteraemic UTIs and likelihood of death. Results 559 bacteraemic episodes occurred in 437 patients. Overall, there were 90 deaths (20.6%) at 7 days and 153 deaths (35.0%) at 30 days. Among patients with catheter-associated bacteraemic UTIs, 7-day and 30-day mortalities associated with each bacteraemic episode were 25/83 (30.1%) and 33/83 (39.8%), respectively. Within this subgroup, the commonest isolates were Escherichia coli, 36 (43.4%), Proteus mirabilis, 11 (13.3%) and Pseudomonas aeruginosa, 9 (10.8%). There were 22 (26.5%) multiple drug-resistant isolates and, of the E coli infections, 6 (16.7%) were extended spectrum β-lactamase producers. In univariate analysis, the variables found to have the strongest association with 7-day mortality were age, Pitt score, Charlson comorbidity index (CCI), medical speciality and site of infection. Compared with CVC-associated bacteraemic infections, there was a significant association between catheter-associated bacteraemic UTIs and 7-day mortality (OR 4.16, 95% CI 1.86 to 9.33). After adjustment for age and CCI, this association remained significant (OR 2.90, 95% CI 1.19 to 7.07). Conclusions Compared with CVC-associated bacteraemic infections, catheter-associated bacteraemic UTIs were significantly associated with 7-day mortality. Efforts to reduce these infections should be prioritised.

Evaluation of two-fold fully conditional specification multiple imputation for longitudinal electronic health record data

Most implementations of multiple imputation (MI) of missing data are designed for simple rectangular data structures ignoring temporal ordering of data. Therefore, when applying MI to longitudinal data with intermittent patterns of missing data, some alternative strategies must be considered. One approach is to divide data into time blocks and implement MI independently at each block. An alternative approach is to include all time blocks in the same MI model. With increasing numbers of time blocks, this approach is likely to break down because of co-linearity and over-fitting. The new two-fold fully conditional specification (FCS) MI algorithm addresses these issues, by only conditioning on measurements, which are local in time. We describe and report the results of a novel simulation study to critically evaluate the two-fold FCS algorithm and its suitability for imputation of longitudinal electronic health records. After generating a full data set, approximately 70% of selected continuous and categorical variables were made missing completely at random in each of ten time blocks. Subsequently, we applied a simple time-to-event model. We compared efficiency of estimated coefficients from a complete records analysis, MI of data in the baseline time block and the two-fold FCS algorithm. The results show that the two-fold FCS algorithm maximises the use of data available, with the gain relative to baseline MI depending on the strength of correlations within and between variables. Using this approach also increases plausibility of the missing at random assumption by using repeated measures over time of variables whose baseline values may be missing.

Midlife Risk Factors for Impaired Physical and Cognitive Functioning at Older Ages: A Cohort Study

Background: Previous studies examined midlife risk factors separately for old-age impaired physical and cognitive functioning. We determined the overlap of risk factors for both domains of functioning within the same setting. Methods: Biological and behavioral risk factors at age 50 years and cognitive and physical functioning were assessed 18 (SD = 5) years later in the Whitehall II study (N = 6,316). Impaired physical functioning was defined as ≥1 limitation on the activities of daily living scale. Impaired cognitive functioning was defined as Mini-Mental State Examination score <27. Two statistical analyses were employed: minimally adjusted analysis (for age, sex, and ethnicity) and mutually adjusted analysis (including all risk factors). Missing data on risk factors were imputed. Results: After confounder adjustment, impaired physical and cognitive functioning at older ages were predicted by hypertension (odds ratios [ORs] 1.80 95% confidence interval [CI] 1.39–2.33 and 1.57 95% CI 1.07–2.31, respectively), poor lung function (1.51 95% CI 1.28–1.78 and 1.31 95% CI 1.08–1.59), and physical inactivity, marginally in the case of cognitive function (1.50 95% CI 1.19–1.90 and 1.27 95% CI 0.99–1.62) at age 50 years. Impaired physical functioning but not cognitive functioning was additionally predicted by depression and higher body mass index (1.72 95% CI 1.46–2.03 and 1.29 95% CI 1.16–1.44, respectively). Conclusions: Several midlife risk factors are associated with impaired physical and cognitive functioning in old age, supporting a unified prevention policy. Analysis of 12 risk factors at age 50 suggests that strategies targeting physical inactivity, hypertension, and poor lung function will reduce impairments in both cognitive and physical functioning in old age.

Two-stage method to remove population- and individual-level outliers from longitudinal data in a primary care database.

In the UK, primary care databases include repeated measurements of health indicators at the individual level. As these databases encompass a large population, some individuals have extreme values, but some values may also be recorded incorrectly. The challenge for researchers is to distinguish between records that are due to incorrect recording and those which represent true but extreme values. This study evaluated different methods to identify outliers.

Detection of Outliers Due to Participants' Non-Adherence to Protocol in a Longitudinal Study of Cognitive Decline

Background Participants’ non adherence to protocol affects data quality. In longitudinal studies, this leads to outliers that can be present at the level of the population or the individual. The purpose of the present study is to elaborate a method for detection of outliers in a study of cognitive ageing. Methods In the Whitehall II study, data on a cognitive test battery have been collected in 1997-99, 2002-04, 2007-09 and 2012-13. Outliers at the 2012-13 wave were identified using a 4-step procedure: (1) identify cognitive tests with potential non-adherence to protocol, (2) choose a prediction model between a simple model with socio-demographic covariates and one that also includes health behaviours and health measures, (3) define an outlier using a studentized residual, and (4) study the impact of exclusion of outliers by estimating the effect of age and diabetes on cognitive decline. Results 5516 participants provided cognitive data in 2012-13. Comparisons of rates of annual decline over the first three and all four waves of data suggested outliers in three of the 5 tests. Mean residuals for the 2012-13 wave were larger for the basic compared to the more complex prediction model (all p<0.001), leading us to use the latter for the identification of outliers. Residuals greater than two standard deviation of residuals identified approximately 7% of observations as being outliers. Removal of these observations from the analyses showed that both age and diabetes had associations with cognitive decline similar to that observed with the first three waves of data; these associations were weaker or absent in non-cleaned data. Conclusions Identification of outliers is important as they obscure the effects of known risk factor and introduce bias in the estimates of cognitive decline. We showed that an informed approach, using the range of data collected in a longitudinal study, may be able to identify outliers.

Socioeconomic inequality in recovery from poor physical and mental health in mid-life and early old age: prospective Whitehall II cohort study

Background Few studies have examined the influence of socioeconomic status on recovery from poor physical and mental health. Methods Prospective study with four consecutive periods of follow-up (1991–2011) of 7564 civil servants (2228 women) recruited while working in London. Health was measured by the Short-Form 36 questionnaire physical and mental component scores assessed at beginning and end of each of four rounds. Poor health was defined by a score in the lowest 20% of the age–sex-specific distribution. Recovery was defined as changing from a low score at the beginning to a normal score at the end of the round. The analysis took account of retirement status, health behaviours, body mass index and prevalent chronic disease. Results Of 24 001 person-observations in the age range 39–83, a total of 8105 identified poor physical or mental health. Lower grade of employment was strongly associated with slower recovery from poor physical health (OR 0.73 (95% CI 0.59 to 0.91); trend P=0.002) in age, sex and ethnicity-adjusted analyses. The association was halved after further adjustment for health behaviours, adiposity, systolic blood pressure (SBP) and serum cholesterol (OR 0.85 (0.68 to 1.07)). In contrast, slower recovery from poor mental health was associated robustly with low employment grade even after multiple adjustment (OR 0.74 (0.59 to 0.93); trend P=0.02). Conclusions Socioeconomic inequalities in recovery from poor physical health were explained to a considerable extent by health behaviours, adiposity, SBP and serum cholesterol. These risk factors explained only part of the gradient in recovery for poor mental health.

Does pattern mixture modelling reduce bias due to informative attrition compared to fitting a mixed effects model to the available cases or data imputed using multiple imputation?: a simulation study

BackgroundInformative attrition occurs when the reason participants drop out from a study is associated with the study outcome. Analysing data with informative attrition can bias longitudinal study inferences. Approaches exist to reduce bias when analysing longitudinal data with monotone missingness (once participants drop out they do not return). However, findings may differ when using these approaches to analyse longitudinal data with non-monotone missingness.MethodsDifferent approaches to reduce bias due to informative attrition in non-monotone longitudinal data were compared. To achieve this aim, we simulated data from a Whitehall II cohort epidemiological study, which used the slope coefficients from a linear mixed effects model to investigate the association between smoking status at baseline and subsequent decline in cognition scores. Participants with lower cognitive scores were thought to be more likely to drop out. By using a simulation study, a range of scenarios using distributions of variables which exist in real data were compared.Informative attrition that would introduce a known bias to the simulated data was specified and the estimates from a mixed effects model with random intercept and slopes when fitted to: available cases; data imputed using multiple imputation (MI); imputed data adjusted using pattern mixture modelling (PMM) were compared. The two-fold fully conditional specification MI approach, previously validated for non-monotone longitudinal data under ignorable missing data assumption, was used. However, MI may not reduce bias because informative attrition is non-ignorable missing. Therefore, PMM was applied to reduce the bias, usually unknown, by adjusting the values imputed with MI by a fixed value equal to the introduced bias.ResultsWith highly correlated repeated outcome measures, the slope coefficients from a mixed effects model were found to have least bias when fitted to available cases. However, for moderately correlated outcome measurements, the slope coefficients from fitting a mixed effects model to data adjusted using PMM were least biased but still underestimated the true coefficients.ConclusionsPMM may potentially reduce bias in studies analysing longitudinal data with suspected informative attrition and moderately correlated repeated outcome measurements. Including additional auxiliary variables in the imputation model may also reduce any remaining bias.

The utility of a blood culture database to identify patients suitable for outpatient parenteral antibiotic treatment

Background The clinical and cost-effectiveness of outpatient parenteral antimicrobial therapy (OPAT) services are well described. We used a blood culture database as a novel approach to case finding and determined its utility in identifying inpatients suitable for OPAT. Methods From December 2012 to November 2013, consecutive adult inpatients with bacteraemia, and those recruited to OPAT, were prospectively studied. Univariate and multivariate logistic regression analysis were used to investigate the association between bacteraemic patient characteristics and OPAT recruitment. Results There were 470 bacteraemic and 134 OPAT patients. The blood culture database identified 22 (16.4%; CI 10.5 to 23.6) additional patients suitable for OPAT, 4.7% (95% CI 3.0% to 7.0%) of the total bacteraemic cohort. 20 (90.9%) of these patients had community-acquired bacteraemia. Bacteraemic patients with urinary tract infections (UTIs), 11/157 (7.0%; 95% CI 3.5% to 12.2%) were most commonly recruited to OPAT and Escherichia coli was the most common blood culture isolate. In the E. coli bacteraemic subgroup, extended-spectrum β-lactamase (ESBL) producers were significantly higher in the OPAT group, compared with the non-OPAT group, 9/11 (81.8%) vs 17/192 (8.9%), p<0.001. Among OPAT patients, there were no deaths within 30 days and no significant difference in relapse rates between bacteraemic and non-bacteraemic patients, 1/22 (4.6%) vs 5/112 (4.5%). In logistic regression analysis, there were no patient characteristics in the bacteraemic cohort that predicted recruitment to OPAT. In a subgroup analysis of patients with Gram-negative bacteraemia, ESBL production was strongly associated with OPAT recruitment, OR 5.85 (95% CI 1.94 to 17.58), p=0.002. Conclusions A blood culture database proved a useful adjuvant to a clinical referral system, particularly for patients with community onset, multidrug resistant UTIs caused by ESBL producing E. coli. All bacteraemic patients recruited to OPAT received treatment safely and had good clinical outcomes.