Quality and variability of osteoporosis data in general practice computer records: implications for disease registers

Ethical Issues and the Electronic Health Record

Ethical issues related to electronic health records (EHRs) confront health personnel. Electronic health records create conflict among several ethical principals. Electronic health records may represent beneficence because they are alleged to increase access to health care, improve the quality of care and health, and decrease costs. Research, however, has not consistently demonstrated access for disadvantaged persons, the accuracy of EHRs, their positive effects on productivity, nor decreased costs. Should beneficence be universally acknowledged, conflicts exist with other ethical principles. Autonomy is jeopardized when patients' health data are shared or linked without the patients' knowledge. Fidelity is breached by the exposure of thousands of patients' health data through mistakes or theft. Lack of confidence in the security of health data may induce patients to conceal sensitive information. As a consequence, their treatment may be compromised. Justice is breached when persons, because of their socioeconomic class or age, do not have equal access to health information resources and public health services. Health personnel, leaders, and policy makers should discuss the ethical implications of EHRs before the occurrence of conflicts among the ethical principles. Recommendations to guide health personnel, leaders, and policy makers are provided.

Can the Use of Bayesian Analysis Methods Correct for Incompleteness in Electronic Health Records Diagnosis Data? Development of a Novel Method Using Simulated and Real-Life Clinical Data

Background: Patient health information is collected routinely in electronic health records (EHRs) and used for research purposes, however, many health conditions are known to be under-diagnosed or under-recorded in EHRs. In research, missing diagnoses result in under-ascertainment of true cases, which attenuates estimated associations between variables and results in a bias toward the null. Bayesian approaches allow the specification of prior information to the model, such as the likely rates of missingness in the data. This paper describes a Bayesian analysis approach which aimed to reduce attenuation of associations in EHR studies focussed on conditions characterized by under-diagnosis.

Methods: Study 1: We created synthetic data, produced to mimic structured EHR data where diagnoses were under-recorded. We fitted logistic regression (LR) models with and without Bayesian priors representing rates of misclassification in the data. We examined the LR parameters estimated by models with and without priors. Study 2: We used EHR data from UK primary care in a case-control design with dementia as the outcome. We fitted LR models examining risk factors for dementia, with and without generic prior information on misclassification rates. We examined LR parameters estimated by models with and without the priors, and estimated classification accuracy using Area Under the Receiver Operating Characteristic.

Results: Study 1: In synthetic data, estimates of LR parameters were much closer to the true parameter values when Bayesian priors were added to the model; with no priors, parameters were substantially attenuated by under-diagnosis. Study 2: The Bayesian approach ran well on real life clinic data from UK primary care, with the addition of prior information increasing LR parameter values in all cases. In multivariate regression models, Bayesian methods showed no improvement in classification accuracy over traditional LR.

Conclusions: The Bayesian approach showed promise but had implementation challenges in real clinical data: prior information on rates of misclassification was difficult to find. Our simple model made a number of assumptions, such as diagnoses being missing at random. Further development is needed to integrate the method into studies using real-life EHR data. Our findings nevertheless highlight the importance of developing methods to address missing diagnoses in EHR data.

Development of the Andalusian Registry of Patients Receiving Community Case Management, for the follow-up of people with complex chronic diseases

RATIONALE, AIMS AND OBJECTIVES

Complex chronic diseases are a challenge for the current configuration of health services. Case management is a service frequently provided for people with chronic conditions, and despite its effectiveness in many outcomes, such as mortality or readmissions, uncertainty remains about the most effective form of team organization, structures and the nature of the interventions. Many processes and outcomes of case management for people with complex chronic conditions cannot be addressed with the information provided by electronic clinical records. Registries are frequently used to deal with this weakness. The aim of this study was to generate a registry-based information system of patients receiving case management to identify their clinical characteristics, their context of care, events identified during their follow-up, interventions developed by case managers and services used.

METHODS

The study was divided into three phases, covering the detection of information needs, the design and its implementation in the health care system, using literature review and expert consensus methods to select variables that would be included in the registry.

RESULTS

A total of 102 variables representing structure, processes and outcomes of case management were selected for their inclusion in the registry after the consensus phase. A web-based registry with modular and layered architecture was designed. The framework follows a pattern based on the model-view-controller approach. In its first 6 months after the implementation, 102 case managers have introduced an average number of 6.49 patients each one.

CONCLUSIONS

The registry permits a complete and in-depth analysis of the characteristics of the patients who receive case management, the interventions delivered and some major outcomes as mortality, readmissions or adverse events.

Analysis of osteoporosis treatment patterns with bisphosphonates and outcomes among postmenopausal veterans

PURPOSE

Adherence and persistence with bisphosphonates are frequently poor, and stopping, restarting, or switching bisphosphonates is common. We evaluated bisphosphonate change behaviors (switching, discontinuing, or reinitiating) over time, as well as fractures and costs, among a large, national cohort of postmenopausal veterans.

METHODS

Female veterans aged 50+ treated with bisphosphonates during 2003-2011 were identified in Veterans Health Administration (VHA) datasets. Bisphosphonate change behaviors were characterized using pharmacy refill records. Patients' baseline disease severity was characterized based on age, T-score, and prior fracture. Cox Proportional Hazard analysis was used to evaluate characteristics associated with discontinuation and the relationship between change behaviors and fracture outcomes. Generalized estimating equations were used to evaluate the relationship between change behaviors and cost outcomes.

RESULTS

A total of 35,650 patients met eligibility criteria. Over 6800 patients (19.1%) were non-switchers. The remaining patients were in the change cohort; at least half displayed more than one change behavior over time. A strong, significant predictor of discontinuation was ≥5 healthcare visits in the prior year (11-23% more likely to discontinue), and discontinuation risk decreased with increasing age. No change behaviors were associated with increased fracture risk. Total costs were significantly higher in patients with change behaviors (4.7-19.7% higher). Change-behavior patients mostly had significantly lower osteoporosis-related costs than non-switchers (22%-118% lower).

CONCLUSIONS

Most bisphosphonate patients discontinue treatment at some point, which did not significantly increase the risk of fracture in this majority non-high risk population. Bisphosphonate change behaviors were associated with significantly lower osteoporosis costs, but significantly higher total costs.

Improving Osteoporosis Management in Primary Care: An Audit of the Impact of a Community Based Fracture Liaison Nurse

Background

Osteoporosis and associated fragility fractures are a major health problem; they are more common in women over 50 years old. Fracture liaison nurses have been widely used in secondary care to promote the recognition of fragility fractures and to promote the use of bone-sparing medication to reduce the risk of recurrent facture.

Objective

Audit the impact of a primary care based fracture liaison nurse on the detection of fragility fractures in people with osteoporosis and their treatment with a bone-sparing medication.

Method

This audit took place in 12 GP practices using ‘before and after’ cross-sectional extractions of anonymised routine data. We report, for females 50–74 years and ≥75 years old, socio-economic deprivation index, the prevalence of osteoporosis, recording of fragility fractures, dual-energy X-ray absorptiometry (DXA), smoking, and body-mass index (BMI) and use of appropriate bone-sparing medication. We used Altman’s test of independent proportions to compare before and after data.

Results

Recording of the diagnosis of osteoporosis increased from 1.5% to 1.7% (p = 0.059); the rate of DXA scans fell (1.8% to 1.4%; p = 0.002); recording of fractures and fragility fractures more than doubled (0.8% to 2.0%; p<0.001 and 0.5% to 1.5%; p<0.001, respectively) with approximate doubling of the recording of smoking, and BMI (p<0.001 level). Fragility fracture recording rose from 8.8% to 15% in females aged 50 to 74, and from 0.8% to 2.3% in people aged ≥75years old (p<0.001). There appeared to be inequity in the service, people who were least deprived were more likely to receive DXA scans and the more deprived to be prescribed bone sparing agents.

Conclusion

A fracture liaison nurse in primary care has been associated with a period of improved management. Liaison nurses based in different parts of the health system should be tested in a prospective trial.

Methods and dimensions of electronic health record data quality assessment: enabling reuse for clinical research

OBJECTIVE

To review the methods and dimensions of data quality assessment in the context of electronic health record (EHR) data reuse for research.

MATERIALS AND METHODS

A review of the clinical research literature discussing data quality assessment methodology for EHR data was performed. Using an iterative process, the aspects of data quality being measured were abstracted and categorized, as well as the methods of assessment used.

RESULTS

Five dimensions of data quality were identified, which are completeness, correctness, concordance, plausibility, and currency, and seven broad categories of data quality assessment methods: comparison with gold standards, data element agreement, data source agreement, distribution comparison, validity checks, log review, and element presence.

DISCUSSION

Examination of the methods by which clinical researchers have investigated the quality and suitability of EHR data for research shows that there are fundamental features of data quality, which may be difficult to measure, as well as proxy dimensions. Researchers interested in the reuse of EHR data for clinical research are recommended to consider the adoption of a consistent taxonomy of EHR data quality, to remain aware of the task-dependence of data quality, to integrate work on data quality assessment from other fields, and to adopt systematic, empirically driven, statistically based methods of data quality assessment.

CONCLUSION

There is currently little consistency or potential generalizability in the methods used to assess EHR data quality. If the reuse of EHR data for clinical research is to become accepted, researchers should adopt validated, systematic methods of EHR data quality assessment.

Use of a collaborative database for epidemiological analyses and professional practice evaluation

RATIONALE

In nephrology, the NEOERICA project assessed the feasibility of the diagnosis scheme based on a general practice database. This approach opened a new area where routinely collected data could be used for purposes other than patient management, such as epidemiological analysis and professional practice evaluation. In Lyon, the TIRCEL network is made up of a coordination team and an online database. In 2008, a total of 468 professionals participated and 983 patients were in the database corresponding to 4114 consultations and 9250 biological assessments.

OBJECTIVE

To investigate the impact of a quality control process on the data from operational databases.

METHODS

We set up a quality control process and we described the impact of this process on data. We also specifically investigated the role of measurement scales in error frequency and we studied the impact of data quality on variables which could be used for professional practice evaluation.

RESULTS

Quality control allowed us to detect as inconsistent data 7.5% of tested data. This rate is linked to the parameters and varied from less than 1% (weight, diastolic blood pressure and urinary sodium) to more than 30% (serum iron and ferritin). Quality control led mainly to the validation of the identified data for 80.4%, a direct correction was realized for 12.9%, 5.6% by the lab and only 1.2% were set to missing. Average proteinuria was modified with the quality control process (2.09 g per 24 hours vs. 0.82 g per 24 hours); however, the median remained stable (0.21 g per 24 hours).

CONCLUSION

Specialty databases such as TIRCEL could not be used for epidemiological research or for the extraction of indicators for professional practice evaluation without strict quality control or the set-up of data-entering limits and alarms.

Automated identification of miscoded and misclassified cases of diabetes from computer records

AIMS

To develop a computer processable algorithm, capable of running automated searches of routine data that flag miscoded and misclassified cases of diabetes for subsequent clinical review.

METHOD

Anonymized computer data from the Quality Improvement in Chronic Kidney Disease (QICKD) trial (n = 942,031) were analysed using a binary method to assess the accuracy of data on diabetes diagnosis. Diagnostic codes were processed and stratified into: definite, probable and possible diagnosis of Type 1 or Type 2 diabetes. Diagnostic accuracy was improved by using prescription compatibility and temporally sequenced anthropomorphic and biochemical data. Bayesian false detection rate analysis was used to compare findings with those of an entirely independent and more complex manual sort of the first round QICKD study data (n = 760,588).

RESULTS

The prevalence of definite diagnosis of Type 1 diabetes and Type 2 diabetes were 0.32% and 3.27% respectively when using the binary search method. Up to 35% of Type 1 diabetes and 0.1% of Type 2 diabetes were miscoded or misclassified on the basis of age/BMI and coding. False detection rate analysis demonstrated a close correlation between the new method and the published hand-crafted sort. Both methods had the highest false detection rate values when coding, therapeutic, anthropomorphic and biochemical filters were used (up to 90% for the new and 75% for the hand-crafted search method).

CONCLUSIONS

A simple computerized algorithm achieves very similar results to more complex search strategies to identify miscoded and misclassified cases of both Type 1 diabetes and Type 2 diabetes. It has the potential to be used as an automated audit instrument to improve quality of diabetes diagnosis.

Using WHO-FRAX to describe fracture risk: experience in primary care

Ideally those at highest risk of fracture should be identified prior to fracture occurrence to reduce mortality, morbidity and costs. Case-finding strategies for those at high risk of first fracture or systematic case-finding strategies following fracture are recommended in the UK, rather than population-based screening to identify individuals at high fracture risk. General practices in the UK hold relevant data on individuals beyond fracture history that could allow identification of a wider group of patients at highest risk of fracture. The aim of the paper is to evaluate the feasibility of applying the WHO-FRAX fracture risk calculator to general practice populations using existing recorded data. A cross-sectional study of 2467 women aged 50 years and older (mean 66.2 years, standard deviation = 11.3) registered with two Scottish General Practices with low deprivation (one semi-rural, one urban) was undertaken. Patient data were extracted from the two general practices' patient information databases and the WHO-FRAX calculator was applied to these data. WHO-FRAX calculation was possible on 1872 patients. Of these, 687 patients were found to have a high fracture risk (risk of major facture ≥15% and or risk of hip fracture ≥3% - 37% of the WHO-FRAX assessed cohort) and should be considered for follow-up. In conclusion, use of the WHO-FRAX calculator using general practice-held data is feasible and can help to identify a patient group at higher fracture risk. Further evaluation and treatments can then be targeted at this group.

Temporal and within practice variability in the health improvement network

BACKGROUND

The Health Improvement Network (THIN) database is a primary care electronic medical record database in the UK designed for pharmacoepidemiologic research. Matching on practice and calendar year often is used to account for secular trends in time and differences across practices. However, little is known about the consistency within practices across observation years and among practices within a given year, in THIN or other large medical record databases.

METHODS

We analyzed mortality rates, cancer incidence rates, prescribing rates, and encounter rates across 415 practices from 2000 to 2007 using a practice-year as the unit of observation in separate random and fixed effects longitudinal Poisson regression models. Adjusted models accounted for aggregate practice-level characteristics (smoking, obesity, age, and Vision software experience).

RESULTS

In adjusted models, subsequent calendar years were associated with lower reported mortality rates, increasing cancer reporting rates, increasing prescriptions per patient, and decreasing encounters per patient, with a corresponding linear trend (p < 0.001 for all analyses). For calendar year 2007, the ratio of the 75th percentile to the 25th percentile for crude rate of cancer, mortality, prescriptions, and encounters was 1.63, 1.63, 1.45, and 1.42, respectively. Adjusting for practice characteristics reduced the among-practice variation by approximately 40%.

CONCLUSIONS

THIN data are characterized by secular trends and among-practice variation, both of which should be considered in the design of pharmacoepidemiology studies. Whether these are trends in data quality or true secular trends could not be definitively differentiated.

Clinical risk factors for fracture among postmenopausal patients at risk for fracture: a historical cohort study using electronic medical record data

Osteoporosis represents a growing health burden, but recognition and screening rates are low. Electronic reminders for osteoporosis have been beneficial but are not based on clinical risk factors. Available risk screening tools may contain useful constructs for creating risk-based electronic medical record (EMR) reminders. Using a cohort study design among women ≥50 years with osteoporosis or osteoporosis risk, we searched the EMR for five World Health Organization (WHO) clinical risk factors including older age, lower body mass index (BMI), low bone mineral density (BMD), history of fracture since age 50, and maternal history of osteoporosis or fracture. Rates of reporting were lower than expected for BMD (6.8%), personal history of fracture (3.5%), and maternal history of fracture (0.3%). Despite the limitations, the EMR data were useful for identifying women at highest risk for fracture. Some evidence of bias in reporting rates was present. EMR data can be useful for identifying high fracture risk patients.

Review: electronic health records and the reliability and validity of quality measures: a review of the literature

Previous reviews of research on electronic health record (EHR) data quality have not focused on the needs of quality measurement. The authors reviewed empirical studies of EHR data quality, published from January 2004, with an emphasis on data attributes relevant to quality measurement. Many of the 35 studies reviewed examined multiple aspects of data quality. Sixty-six percent evaluated data accuracy, 57% data completeness, and 23% data comparability. The diversity in data element, study setting, population, health condition, and EHR system studied within this body of literature made drawing specific conclusions regarding EHR data quality challenging. Future research should focus on the quality of data from specific EHR components and important data attributes for quality measurement such as granularity, timeliness, and comparability. Finally, factors associated with poor or variability in data quality need to be better understood and effective interventions developed.

Ethical issues and the electronic health record

Ethical issues related to electronic health records (EHRs) confront health personnel. Electronic health records create conflict among several ethical principals. Electronic health records may represent beneficence because they are alleged to increase access to health care, improve the quality of care and health, and decrease costs. Research, however, has not consistently demonstrated access for disadvantaged persons, the accuracy of EHRs, their positive effects on productivity, nor decreased costs. Should beneficence be universally acknowledged, conflicts exist with other ethical principles. Autonomy is jeopardized when patients' health data are shared or linked without the patients' knowledge. Fidelity is breached by the exposure of thousands of patients' health data through mistakes or theft. Lack of confidence in the security of health data may induce patients to conceal sensitive information. As a consequence, their treatment may be compromised. Justice is breached when persons, because of their socioeconomic class or age, do not have equal access to health information resources and public health services. Health personnel, leaders, and policy makers should discuss the ethical implications of EHRs before the occurrence of conflicts among the ethical principles. Recommendations to guide health personnel, leaders, and policy makers are provided.

Using computers to identify non-compliant people at increased risk of osteoporotic fractures in general practice: a cross-sectional study

BACKGROUND

National guidelines recommend bisphosphonates for secondary prevention of osteoporotic fractures; however, poor compliance may result in sub-optimal prevention.

OBJECTIVE

This study reports the feasibility of using GP electronic records to identify poorly compliant post-menopausal women who may be at increased risk of fragility fractures.

DESIGN

Cross-sectional study of general practice computer records.

SUBJECTS

Women over 45 years, registered in 29 practices across England with a total population of approximately 200,000.

METHODS

MIQUEST (Morbidity Information Query and Export Syntax) a data extraction application was used to extract prescription, diagnostic data and probable fragility fractures (hip, vertebral, wrist). All women >45 years who received a first prescription for a weekly bisphosphonate (alendronate or risedronate) at least a year before data extraction were identified. Each record was examined to determine the number of days of prescribed treatment.

RESULTS

Of 97992 registered women, 44% (42734) were >45 years. Prevalence of likely fragility fractures in women over 45 was 5.1% (2195/42734). 3.0% (1286/42734, mean age 72 years) received a prescription for a bisphosphonate in the 360 day period prior to data extraction with a median duration of treatment of 267 days. 45% (584/1286) received prescriptions covering >288/360 days (Medicine Possession Ratio >80%); 13% (161/1286) collected prescriptions covering >360 days. In those prescribed bisphosphonates, 23% (294/1286) had a likely fragility fracture.

CONCLUSIONS

Women >45 years with probable fragility fractures are more likely to be prescribed bisphosphonates, though less than half will be actually taking them as prescribed. GPs should use computer technology to identify poorly compliant patients who are unnecessarily at risk of fracture.