Matching prescription claims with medication data for nursing home residents: implications for prescriber feedback, drug utilisation studies and selection of prescription claims database

Patterns of high-risk prescribing and other factors in relation to receipt of a home medicines review: a prospective cohort investigation among adults aged 45 years and over in Australia

OBJECTIVES

To quantify the relationship between home medicines review (HMR) receipt in older adults and sociodemographic, medication-related and health factors.

DESIGN

Prospective cohort analysis.

SETTINGS, PARTICIPANTS, MEASUREMENTS

Questionnaire data from a population-based cohort study of individuals aged ≥45 years, Sydney, Australia were linked with primary healthcare data, medication and hospitalisation data, to ascertain factors associated with HMR receipt during the period July 2009-June 2014. Medication-related factors included exposure to five and more medications (polypharmacy), narrow therapeutic index medicines, potentially inappropriate prescribing defined using Beers Criteria medicines, and anticholinergic and sedative drugs, defined using the Drug Burden Index (DBI). Poisson and Cox regression models were used to evaluate HMR receipt in relation to sociodemographic, behavioural and health characteristics, and time-varying factors including medication use and hospitalisations.

PRIMARY OUTCOME

HMR receipt during the 5-year study period.

RESULTS

Over 5 years of follow-up, 4.7% (n=6115) of 131 483 participants received at least one HMR. Five-year HMR receipt was: 1.5% in people using <5 medications at baseline, 6.8% with 5-9 medications, 12.7% with ≥10 medications, 8.8% using Narrow Therapeutic Index medicines, 6.8% using Beers Criteria potentially inappropriate medicines and 7.4% using DBI medicines. Age-sex stratified HRs for HMR receipt were 6.07 (95% CI: 5.58 to 6.59) and 12.46 (11.42 to 13.59) for concurrent use of 5-9 and ≥10 versus <5 medications, respectively. The age-sex adjusted rate ratio for HMR receipt was 2.65 (2.51 to 2.80) with poor versus good self-reported health; this association was attenuated substantially following additional adjustment for polypharmacy.

CONCLUSIONS

HMR was common in individuals using multiple medications, a formal indication for general practitioner referral and, to a lesser extent, with poorer health and other markers of high-risk prescribing. Despite this, HMR use over a 5-year period was generally below 10%, even in high-risk groups, suggesting substantial potential for improvement in uptake and targeting.

A systematic review of methods for determining cross-sectional active medications using pharmacy databases

PURPOSE

Pharmacy dispensing databases are often used to identify patients' medications at a particular time point, for example to measure prescribing quality or the impact of medication use on clinical outcomes. We performed a systematic review of studies that examined methods to assess medications in use at a specific point in time.

METHODS

Comprehensive literature search to identify studies that compared active medications identified using pharmacy databases to medications identified using nonautomated data sources. Two investigators independently reviewed abstracts and full-text material.

RESULTS

Of 496 studies screened, 29 studies evaluating 50 comparisons met inclusion criteria. Twenty-nine comparisons evaluated fixed look-back period approaches, defining active medications as those filled in a specified period prior to the index date (range 84-730 days). Fourteen comparisons evaluated medication-on-hand approaches, defining active medications as those for which the most recent fill provided sufficient supply to last through the study index date. Sensitivity ranged from 48% to 93% for fixed look-back period approaches and 35% to 97% for medication-on-hand approaches. Interpretation of comparative performance of methods was limited by use of different reference sources, target medication classes, and databases across studies. In four studies with head-to-head comparisons of these methods, sensitivity of the medication-on-hand approach was a median of 7% lower than the corresponding fixed look-back approach.

CONCLUSIONS

The reported accuracy of methods for identifying active medications using pharmacy databases differs greatly across studies. More direct comparisons of common approaches are needed to establish the accuracy of methods within and across populations, medication classes, and databases.

Antipsychotic polypharmacy in older Australians

ABSTRACTBackground:Antipsychotics are commonly used, and the rate of use is highest, among those aged 65 years or over, where the risk of adverse events is also high. Up to 20% of younger adults use more than one antipsychotic concurrently; however there are few studies on the prevalence of antipsychotic polypharmacy in older people. We aimed to analyze antipsychotic use in elderly Australians, focusing on the prevalence of antipsychotic polypharmacy and the use of medicines to manage adverse events associated with antipsychotics.

METHODS

A cross-sectional study was conducted using Australian Department of Veterans' Affairs (DVA) administrative claims data for the period 1 March 2014 to 30 June 2014. Veterans dispensed at least one antipsychotic medicine during the study period was included. We determined the number of participants dispensed antipsychotic polypharmacy and the number of participants dispensed medicines to manage antipsychotic side effects.

RESULTS

There were 7,412 participants with a median age of 86 years. Fifty-one percent (n=3,784) were women and 48% (n=3,569) lived in residential aged-care. Fifty one participants (0.7%) were dispensed anticholinergic medicines indicated for the management of antipsychotic-associated extrapyramidal movement disorders and eight (0.1%) were dispensed medicines for the management of hyperprolactinemia. Five percent of participants (n=365) received dual antipsychotics. Dual antipsychotic users were more likely to be under the care of a psychiatrist or to have had a mental health hospitalization than those using a single antipsychotic.

CONCLUSIONS

Antipsychotic polypharmacy occurred in one in 20 elderly persons, indicating that there is room for improvement in antipsychotic use in elderly patients.

Medicines taken by older Australians after transient ischaemic attack or ischaemic stroke: a retrospective database study

BACKGROUND

Guidelines recommend patients diagnosed with transient ischaemic attack (TIA) or ischaemic stroke receive antihypertensive, antithrombotic and lipid lowering medicines. Reassessment of the need for medicines associated with an increased risk of stroke is also recommended.

OBJECTIVE

To determine changes in the use of medicines recommended for secondary stroke prevention, medicines commonly used for treating stroke-related complications and medicines not recommended for use after ischaemic stroke, and to determine patient characteristics associated with use of all three stroke prevention medicines after TIA or ischaemic stroke. Setting Administrative health claims data from the Australian Government Department of Veterans' Affairs.

METHOD

This retrospective study included patients with a first-ever hospitalisation for TIA or ischaemic stroke in 2009 and alive at 4 months after discharge. Changes to medicines dispensed in the 4 months before and after hospitalisation were compared using McNemar's test. Log binomial regression analysis was used to determine patient characteristics associated with use of all three secondary stroke prevention medicines after hospitalisation for TIA or ischaemic stroke.

MAIN OUTCOME MEASURE

Prevalence of medicine use after hospitalisation.

RESULTS

1541 patients (853 TIA, 688 ischaemic stroke) were included, with a median age of 85 years. High use of antihypertensive (82% TIA, 86 % ischaemic stroke) and antithrombotic (84% TIA, 90% ischaemic stroke) medicines was observed postdischarge, with 58% of TIA and 73% of ischaemic stroke patients receiving lipid lowering therapy. Half of the population (47% TIA, 61% ischaemic stroke) were dispensed all three classes of medicines recommended for secondary stroke prevention after discharge. Ischaemic stroke patients, younger patients, patients with more comorbid conditions and those discharged home were more likely to receive all three recommended medicine classes. Antibiotics (45% TIA, 46% ischaemic stroke), paracetamol (44% TIA, 47% ischaemic stroke), antidepressants (26% TIA, 31% ischaemic stroke) and laxatives (24% TIA, 32% ischaemic stroke) were commonly used after discharge. Increased use of sedatives and reduced use of non-steroidal anti-inflammatories was also observed after discharge.

CONCLUSION

Changes to pharmacotherapy after TIA or ischaemic stroke were consistent with treatment for stroke risk factors and common stroke-related complications. Use of secondary stroke prevention medicines may be further improved among TIA patients.

Antithrombotic use following transient ischaemic attack or ischaemic stroke among older Australians with atrial fibrillation

Hospital audits may underestimate anticoagulant use among acute ischaemic stroke patients with atrial fibrillation (AF), as treatment may commence after discharge. To account for this, antithrombotic use in the 4 months after hospitalisation for transient ischaemic attack or ischaemic stroke among AF patients was assessed using claims data. Results suggest that treatment may be commenced soon after discharge and should be considered when assessing prevalence of use.

Prevalence of preventable medication-related hospitalizations in Australia: an opportunity to reduce harm

OBJECTIVE

To identify the prevalence of potentially preventable medication-related hospitalizations amongst elderly Australian veterans by applying clinical indicators to administrative claims data.

DESIGN AND SETTING

Retrospective cohort study in the Australian veteran population from 1 January 2004 to 31 December 2008.

PARTICIPANTS

A total of 109 044 veterans with one or more hospitalizations defined by the medication-related clinical indicator set, during the 5-year study period.

MAIN OUTCOME MEASURE

The prevalence of potentially preventable medication-related hospitalizations as a proportion of all hospitalizations defined by the clinical indicator set.

RESULTS

During the 5-year study period, there were a total of 1 630 008 hospital admissions of which 216 527 (13.3%) were for conditions defined by the medication-related clinical indicator set for 109 044 veterans. The overall proportion of potentially preventable medication-related hospitalizations was 20.3% (n= 43 963). Of the 109 044 veterans included in the study, 28 044 (25.7%) had at least one potentially preventable medication-related hospitalization and 7245 (6.6%) veterans had two or more potentially preventable admissions. Conditions with both a high prevalence of hospitalization and preventability included asthma/chronic obstructive pulmonary disorder, depression and thromboembolic cerebrovascular event (23.3, 18.5 and 18.3%, respectively, were potentially preventable). Other hospitalizations that were less common but had a high level of preventability (at least 20%) included hip fracture, impaction, renal failure, acute confusion, bipolar disorder and hyperkalaemia.

CONCLUSIONS

The results of this study highlight those conditions where hospitalizations could potentially be avoided through improved medication management. Strategies to increase the awareness, identification and resolution of these medication-related problems contributing to these hospitalizations are required in Australia.

Managing glaucoma in those with co-morbidity: not as easy as it seems

PURPOSE

To identify the extent of use of medicines recommended to be used with caution in glaucoma patients with specified comorbidities and to determine evidence of associated harm. Retrospective cohort analysis from administrative claims data and prescription/event sequence symmetry analysis.

PARTICIPANTS

Australian Government Department of Veterans' Affairs treatment card holders dispensed glaucoma eye-drops.

MAIN OUTCOME MEASURES

Proportion of veterans with glaucoma and diabetes, airways disease, heart failure, ischemic heart disease or depression, dispensed glaucoma eye drops which should be used with caution. For harms, outcome measures were hospitalizations for airways disease and heart disease.

RESULTS

The cohort analysis included 25,984 veterans. Of these, 88% with airways disease were dispensed glaucoma eye drops with the potential to aggravate airways disease, 43% with heart failure were dispensed topical beta-blockers and 49% with depression received glaucoma eye drops which should be used cautiously in those with depression. We found increased risk of initiation of inhaled beta-agonist following timolol (adjusted sequence ratio (ASR) 1.48, 99% CI 1.22-1.78) and latanoprost (ASR 1.24, 99% CI 1.11-1.38) initiation. We found increased risk of inhaled corticosteroid initiation following initiation of timolol (ASR 1.43, 99% CI 1.13-1.81). There was increased risk of antidepressant initiation following timolol initiation (ASR 1.24, 99% CI 1.07-1.43), and latanoprost (ASR 1.16, 99% CI 1.03-1.31). There was also increased risk of hospitalization for bradycardia following timolol initiation (ASR 2.22,99% CI 1.15-4.31).

CONCLUSION

Use of glaucoma eye drops recommended to be used with caution in co-morbidities is common and was associated with adverse outcomes. Awareness of co-morbidities is required in the selection and prescription of glaucoma eye drops.

Pharmaceutical brand substitution in Australia: Identifying factors associated with having multiple brand substitutions

Objective

The aim was to determine the frequency with which people have multiple brand changes for more than one medicine and to identify factors associated with having multiple brand substitutions.

Method

The setting was the Repatriation Pharmaceutical Benefits Scheme, a national subsidised scheme for medicines supply in Australia. We used a retrospective cohort design using prescription claims data for subsidised medicine dispensings from 1 June 2005 to 31 August 2006. Analysis was limited to patients with the opportunity for brand substitution of two or more medicines. ‘Switches’ were identified for each medicine if different brand or generic products were supplied at consecutive dispensings. Multivariate analysis was conducted to identify factors associated with patients who had multiple switches during follow-up.

Key findings

A total of 84 040 people were included. On average, they received 11 prescription medicines. Forty-nine per cent of people received the same product throughout follow-up for each medicine and 34% had a single brand substitution. Seventeen per cent had multiple switches for one or more medicine; however, only 3% of all patients had multiple switches for more than one medicine. Independent factors associated with having multiple switches were increasing number of hospital admissions, prescription medicines, co-morbidities, prescribers, dispensing pharmacies and living in an aged-care facility or city.

Conclusions

Most patients do not have multiple brand substitutions of their medicine, even when all medicine use is considered. This is the first study to identify factors associated with having multiple brand substitutions. Quality use of medicines interventions targeting individuals with these risk factors could minimise the potential for patient confusion as a result of multiple brand changes.

Good research practices for comparative effectiveness research: approaches to mitigate bias and confounding in the design of nonrandomized studies of treatment effects using secondary data sources: the International Society for Pharmacoeconomics and Outcomes Research Good Research Practices for Retrospective Database Analysis Task Force Report--Part II

OBJECTIVES

The goal of comparative effectiveness analysis is to examine the relationship between two variables, treatment, or exposure and effectiveness or outcome. Unlike data obtained through randomized controlled trials, researchers face greater challenges with causal inference with observational studies. Recognizing these challenges, a task force was formed to develop a guidance document on methodological approaches to addresses these biases.

METHODS

The task force was commissioned and a Chair was selected by the International Society for Pharmacoeconomics and Outcomes Research Board of Directors in October 2007. This report, the second of three reported in this issue of the Journal, discusses the inherent biases when using secondary data sources for comparative effectiveness analysis and provides methodological recommendations to help mitigate these biases.

RESULTS

The task force report provides recommendations and tools for researchers to mitigate threats to validity from bias and confounding in measurement of exposure and outcome. Recommendations on design of study included: the need for data analysis plan with causal diagrams; detailed attention to classification bias in definition of exposure and clinical outcome; careful and appropriate use of restriction; extreme care to identify and control for confounding factors, including time-dependent confounding.

CONCLUSIONS

Design of nonrandomized studies of comparative effectiveness face several daunting issues, including measurement of exposure and outcome challenged by misclassification and confounding. Use of causal diagrams and restriction are two techniques that can improve the theoretical basis for analyzing treatment effects in study populations of more homogeneity, with reduced loss of generalizability.

Older Australians' medication use: self-report by phone showed good agreement and accuracy compared with home visit

OBJECTIVE

To ascertain the accuracy of telephone-interview method for measuring older people's medication use ("self-report by phone") by determining agreement between results from this method and from a home visit ("home inventory").

STUDY DESIGN AND SETTING

An agreement study involving community-dwelling patients aged more than 65 years, selected from four general practices in the Hunter Region of Australia. Commonly used classes of drugs were selected for comparison.

RESULTS

Of 154 patients, 14 participants were ineligible, because they had hearing problems (9) or did not use any medicines (5). The response rate was 70% (98 of 140). The observed overall agreement and prevalence-adjusted and bias-adjusted kappa coefficients were very high for all prescribed drug categories, but lower for over-the-counter (OTC) and complementary medicines. Specificity of the self-report by phone compared with home inventory was consistently high across all drug classes. Sensitivity values were more than 89% for all drug classes but were lower for OTC and complementary medicines (74%) and paracetamol (78%). Similar patterns were found for negative predictive values. Positive predictive values were lower for drugs used on an as-needed basis.

CONCLUSION

Measuring patient's medication use by telephone is an accurate and relatively inexpensive alternative to home-inventory methods, and has merit for use in future studies of older patients' drug use.

Accuracy of telephone self-report of drug use in older people and agreement with pharmaceutical claims data

OBJECTIVE

To determine the agreement between two measures of medication use, namely telephone interview self-report and pharmaceutical claims data, in an elderly population.

METHODS

An agreement study of 566 community-dwelling, general practice patients aged > or =65 years was conducted to compare self-reported use of medicines with pharmaceutical claims data for different retrieval periods. Classes of drugs commonly used in the elderly were selected for comparison.

RESULTS

1094 people were eligible for the main study. Of these, 697 people completed a follow-up survey and 625 of these patients consented to the release of pharmaceutical claims data. A further 59 participants were excluded from the analysis because they had a home visit instead of a telephone interview. The proportion of observed agreement between the telephone self-report and the various retrieval periods was consistently high. Kappa coefficients showed good to very good agreement (> or = 0.75) with retrieval periods of 30, 60 and 90 days for benzodiazepines, low-risk NSAIDs, thiazide diuretics and most other drugs. The specificity of self-reported medication use compared with claims data was consistently high across all drug classes, suggesting that people usually did not mention drugs that were not included in the claims data. Sensitivity values varied according to drug class and retrieval period, and were lower for NSAIDs than for benzodiazepines and thiazide diuretics. Decline in sensitivity with increased retrieval periods was most marked for benzodiazepines, NSAIDs and low-risk NSAIDs, which are often used on an as-needed basis. Positive predictive values increased with longer retrieval periods

CONCLUSION

High agreement and accuracy were demonstrated for self-reported use of medicines when patients were interviewed over the telephone compared with pharmaceutical claims data. The telephone inventory method can be used in future studies for accurately measuring drug use in older people when claims data are not available.

Diabetes processes of care in the Australian veteran population

OBJECTIVE

To determine the proportion of veterans with diabetes who had claims for recommended processes of diabetes care in 2005.

DESIGN

Observational study of 19,810 veterans' dispensed medicines indicative of diabetes. Claims to the Department of Veterans' Affairs for HbA(1c), microalbuminuria, ophthalmology or optometry services, podiatry services, diabetes care plans, medication reviews, case conferences and GP management plans were assessed.

RESULTS

Overall 63% had at least one HbA(1c) claim in the year under review, while 40% had a microalbuminuria claim. The majority (87%) had a claim for an ophthalmology or optometry service, while two-thirds had a podiatry service claim. Only one-fifth had claims for an annual diabetes care plan. Just under half had a claim for any type of care plan, including medication review, discharge plan, case conference, GP management plan or health care plan. Veterans resident in aged-care facilities were significantly less likely to have claims for any of these services apart from medication review.

CONCLUSIONS

While administrative claims data may under-represent some processes of care carried out routinely in general practice, this study suggests under-utilisation of services for veterans with diabetes, particularly those resident in aged-care facilities.

The influence of the Pharmaceutical Benefits Scheme (PBS) on inappropriate prescribing in Australian nursing homes

OBJECTIVES

To determine the prevalence of inappropriate prescribing, defined by applying modified Beers' criteria, and to examine the influence of the Pharmaceutical Benefits Scheme (PBS), Australia's national scheme for subsidising medicines, on inappropriate prescribing.

METHODS

Cross-sectional survey of nursing home records, including 7-days data from medication charts.

SETTING

Fiveteen randomly selected nursing homes (998 residents) in Southeast Queensland and Northern New South Wales, Australia.

MAIN OUTCOME MEASURES

The prevalence of inappropriate prescribing as defined by modified Beers' criteria and its correlation with PBS restrictions.

RESULTS

18.5% of residents were ordered one or more inappropriate medications, and 1.5% of residents were ordered two or more. The level of PBS restriction and the percentage of residents ordered a medication were highly correlated (p = -0.87, P < 0.001). Medications in Beers' criteria that were not listed (subsidised) on the PBS were not ordered for any resident. PBS medicines with subsidies restricted to certain populations or indications were ordered for 0% to 0.1% of residents. Dextropropoxyphene, diazepam, amitriptyline and methyldopa were the only medications in Beers' criteria prescribed to more than 0.5% of residents. Dextropropoxyphene was only subsidised for war veterans, with a caution warning of its potential to cause drug dependence, while diazepam, amitriptyline and methyldopa were listed on the PBS without any subsidy restrictions.

CONCLUSION

Increases in the level of PBS restriction were associated with decreases in the prevalence of inappropriate prescribing, The targeting of drug subsidies to reduce inappropriate prescribing warrants further investigation.

Reliability of medicaid claims versus medical record data: in a cost analysis of palivizumab

BACKGROUND

Palivizumab is one of the most costly paediatric medications for Medicaid and other health plans. It is uncertain whether the costs of palivizumab administration are justified in specific risk groups. Ongoing investigations of palivizumab require identification of reliable data sources.

OBJECTIVE

To estimate the reliability between Medicaid paid claims data and medical records documentation in a cost analysis of palivizumab.

STUDY DESIGN

A cross-sectional study of data reliability was performed within a cost analysis study of palivizumab in a cohort of at-risk patients during the US 2002/3 respiratory syncytial virus season. Exposure classification (to palivizumab prophylaxis) was compared using Medicaid claims and medical records data.

METHODS

The study was performed in 28 widely dispersed paediatric medical practices serving North Carolina, USA Medicaid patients within the AccessCare/Community Care of North Carolina (CCNC) Program, an enhanced primary care case management programme. Patients were eligible if they were born between 1 March 2002 and 28 February 2003 at 32-35 weeks estimated gestational age, were enrolled in the North Carolina primary care case management Medicaid programme during the study period and were patients of the participating practices. Medicaid healthcare claims were obtained in December 2003 for services provided between October 2002 and May 2003. Medical records were abstracted by community-based case managers. Primary variables included frequency, dates and dose of palivizumab injections. The main outcomes measures were agreement in the number of palivizumab injections, dates of administration and doses of palivizumab between Medicaid paid claims and medical record data.

RESULTS

Injection frequencies matched between medical record and Medicaid claims data for only 46.2% of study participants. Congruence in injection service dates occurred between data sources for only 1% of injections. Doses were similar between data sources for 81.9% of injections.

CONCLUSIONS

In Medicaid recipients receiving palivizumab injection, Medicaid claims data were inconsistent with medical records data. Use of multiple data sources and validation are recommended to identify temporal relationships between drug administration and endpoints of interest.

Medication reviews in the community: results of a randomized, controlled effectiveness trial

AIMS

To examine the effectiveness of a multidisciplinary service model delivering medication review to patients at risk of medication misadventure in the community.

METHODS

The study was carried out in three Australian states; Queensland, New South Wales and Western Australia, and conducted as a randomized, controlled effectiveness trial with the general practitioner (GP) as the unit of randomization. In total, 92 GPs, 53 pharmacists and 400 patients enrolled in the study. The multidisciplinary service model consisted of GP education, patient home visits, pharmacist medication reviews, primary healthcare team conferences, GP implementation of action plans in consultation with patients, and follow-up surgery visits for monitoring. Effectiveness was assessed using the four clinical value compass domains of (i) functional status, (ii) clinical outcomes, (iii) satisfaction and (iv) costs. The domains of functional status (assessed by the health-related quality of life measure SF-36 subscales) and clinical outcomes (as assessed by adverse drug events (ADEs), number of GP visits, hospital services and severity of illness) were measured at baseline and endpoint. Satisfaction was measured by success in implementation and by participant satisfaction at endpoint, and costs (as assessed using medication and healthcare service costs, less intervention costs) were measured preintervention and during the trial. In addition, process evaluation was conducted for intervention patients, in which problems and recommendations from the medication reviews were described.

RESULTS

The model was successfully implemented with 92% of intervention GPs suggesting that the model had improved the care of participating patients, a view shared by 94% of pharmacists. In addition, positive trends in clinical outcomes (ADEs and severity of illness) and costs (an ongoing trend towards reduction in healthcare service costs) were evident, although the trial was limited to a 6-month intervention time. No differences between intervention and control groups were identified for the health-related quality of life domain. The cost-effectiveness ratio for the intervention based on cost savings, reduced adverse events and improved health outcomes was small. The most common problems identified in the medication reviews were potential adverse drug reactions, suboptimal monitoring and adherence/lack of concordance issues. In total, 54.4% of recommendations were enacted, and 23.9% were implemented precisely as recommended in the medication review. Follow-up evaluation showed that 70.9% of actions had a positive outcome, 15.7% no effect and 3.7% had a negative outcome.

CONCLUSIONS

Most studies emphasize efficacy and the best achievable clinical outcomes rather than whether an intervention will be effective in practice. The current trial showed that three of the four domains in the clinical value compass showed trends of improvement or were indeed improved in the relatively short follow-up period of the trial, suggesting that a service based on this model could achieve similar benefits in practice. A domiciliary medication review programme similar to this model has now been implemented into national Australian practice, where GPs and pharmacists are reimbursed by the Australian government for the provision of these services.