High-Cost Users of Prescription Drugs: A Population-Based Analysis from British Columbia, Canada

Predicting high health-cost users among people with cardiovascular disease using machine learning and nationwide linked social administrative datasets

OBJECTIVES

To optimise planning of public health services, the impact of high-cost users needs to be considered. However, most of the existing statistical models for costs do not include many clinical and social variables from administrative data that are associated with elevated health care resource use, and are increasingly available. This study aimed to use machine learning approaches and big data to predict high-cost users among people with cardiovascular disease (CVD).

METHODS

We used nationally representative linked datasets in New Zealand to predict CVD prevalent cases with the most expensive cost belonging to the top quintiles by cost. We compared the performance of four popular machine learning models (L1-regularised logistic regression, classification trees, k-nearest neighbourhood (KNN) and random forest) with the traditional regression models.

RESULTS

The machine learning models had far better accuracy in predicting high health-cost users compared with the logistic models. The harmony score F1 (combining sensitivity and positive predictive value) of the machine learning models ranged from 30.6% to 41.2% (compared with 8.6-9.1% for the logistic models). Previous health costs, income, age, chronic health conditions, deprivation, and receiving a social security benefit were among the most important predictors of the CVD high-cost users.

CONCLUSIONS

This study provides additional evidence that machine learning can be used as a tool together with big data in health economics for identification of new risk factors and prediction of high-cost users with CVD. As such, machine learning may potentially assist with health services planning and preventive measures to improve population health while potentially saving healthcare costs.

Do not PIMP my nursing home ride! The impact of Potentially Inappropriate Medications Prescribing on residents' emergency care use

Nursing home residents often are poly-medicated, which increases their risks of receiving potentially inappropriate medications. This problem has become a major public health issue in many countries, and in particular in France. Indeed, high uses of potentially inappropriate medication prescriptions can lead to adverse effects that are likely to increase emergency room (ER) visits. However, there is a lack of empirical evidence on the causal relationship between the amount of use of potentially inappropriate medications and ER visit risks among nursing homes residents. Indeed, this question is subject to endogeneity issues due to omitted variables that simultaneously affect inappropriate medications prescriptions and ER use. We take advantage of the IDEM Randomized Clinical Trial (Systematic Dementia Screening by Multidisciplinary Team Meetings in Nursing Homes for Reducing Emergency Department Transfers) to overcome that issue. Indeed, randomization in the IDEM intervention group created exogenous variations in potentially inappropriate prescriptions, and was thus used as an instrument. Using an instrumental variable model, we show that over a 12-month period, a 1% increase in the share of potentially inappropriate medications spending in total medication spending leads to a 5.7 percentage point increase in residents' ER use risks (p < 0.001). This effect is robust to various model specifications. Moreover, the intensity of this correlation persists over an 18-month period. While tackling wasteful spending has become a priority in most countries, our results have important policy implications. Indeed, reducing potentially inappropriate medication spending in nursing homes should be a key component of value-based aging policies, which objectives are to reduce inefficient care, and provide health care services centered in people's interest.

High-Cost Users of Prescription Drugs: National Health Insurance Data from South Korea

IMPORTANCE

In OECD countries, pharmaceutical spending reached around 800 billion USD in 2013, accounting for about 20% of total spending in the retail sector. Pharmaceutical expenditures are steadily increasing in South Korea, necessitating strategies to promote efficiency.

OBJECTIVE

This study investigated factors associated with high-cost users (HCUs), who account for the majority of outpatient prescriptions in the total South Korean population. The top 20 frequently prescribed therapeutic subgroups were also investigated.

DESIGN

This is an observational study performed using health insurance claims data in 2019.

PARTICIPANTS

In total, 44,744,632 people (including 6,806,339 aged 65 years or older) who were prescribed outpatient medications were included.

MAIN MEASURES

HCUs were defined as those for whom prescription drug costs were in the top 5%. Multivariate logistic regression analysis was performed using factors including age, insurance type, number of prescription drugs, outpatient visit days, prescription treatment days, and chronic diseases.

RESULTS

HCUs accounted for 3.6 million (5% of the total population) and 1.4 million (21.1% of those 65 years or older). Furthermore, 4.1% of HCUs in the total population had few comorbidities. Male sex, older age, insurance (Medical Aid), comorbidities, chronic diseases, number of prescription drugs, outpatient visit days, and prescription days were all associated with an increased probability of being an HCU. The highest spending was found for B01 (antithrombotic agents) with 0.4 billion USD, followed by C10 (lipid-modifying agents) and A10 (drugs used in diabetes). The proportion of spending for HCUs among the general population was highest in L01 (antineoplastic agents), at 98.2%, and L04 (immunosuppressants), at 87.8%, whereas among the elderly, the highest proportions were found for B01 (antithrombotic agents), at 44.5%, and N06 (antidepressants), at 44.3%.

CONCLUSION

Age and multiple chronic conditions were strongly associated with HCUs, and it seems necessary to reduce drug prescriptions in patients without complex comorbidities. Several measures should target those without multiple chronic conditions who are nonetheless HCUs.

Application of Bayesian network and regression method in treatment cost prediction

Charging according to disease is an important way to effectively promote the reform of medical insurance mechanism, reasonably allocate medical resources and reduce the burden of patients, and it is also an important direction of medical development at home and abroad. The cost forecast of single disease can not only find the potential influence and driving factors, but also estimate the active cost, and tell the management and reasonable allocation of medical resources. In this paper, a method of Bayesian network combined with regression analysis is proposed to predict the cost of treatment based on the patient's electronic medical record when the amount of data is small. Firstly, a set of text-based medical record data conversion method is established, and in the clustering method, the missing value interpolation is carried out by weighted method according to the distance, which completes the data preparation and processing for the realization of data prediction. Then, aiming at the problem of low prediction accuracy of traditional regression model, this paper establishes a prediction model combined with local weight regression method after Bayesian network interpretation and classification of patients' treatment process. Finally, the model is verified with the medical record data provided by the hospital, and the results show that the model has higher prediction accuracy.

Comparison of claims from high-drug cost beneficiaries in Ontario, Canada, and Australia: a cross-sectional analysis

BACKGROUND

Globally, payers are struggling with rising drug costs, driven primarily by the increasing number of high-cost medications used by their beneficiaries. We aimed to compare the annual drug spending on claims from high-drug cost beneficiaries in the province of Ontario, Canada, and Australia.

METHODS

We conducted a cross-sectional analysis of public drug claims in Ontario and Australia from fiscal years 2006 to 2017. We identified the total government costs for prescribed medications per beneficiary. During the study period, public drug coverage in Ontario was provided to all residents 65 years of age and older, those with financial needs, and those living in long-term care or in need of home care. Australia maintains a publicly funded, universal system covering all citizens. Based on annual spending, we divided beneficiaries into 4 cost groups, representing the top 1%, top 5%, top 10% and the remaining 90%. We reported the following for each cost group: medication cost and proportion of total government spending, number of unique drugs dispensed per person and the top 10 most costly drug classes.

RESULTS

In Ontario and Australia, the top 1% of beneficiaries accounted for a large and increasing proportion of all government drug costs, growing from 12% ($405 946 197) to 24% ($1 345 977 248) in Ontario, and from 14% ($86 565 586) to 34% ($416 097 984) in Australia between 2006 and 2017. The most costly drug classes among high-drug cost beneficiaries in both jurisdictions were biologics and hepatitis C treatments.

INTERPRETATION

In both Ontario and Australia, a small number of beneficiaries accounted for a large proportion of public drug spending, driven largely by the use of expensive medications. The current development of potential national pharmacare strategies in Canada must optimize the use of high-cost drugs to ensure the sustainability of the program.

Regional variation in potentially inappropriate medicine use in older adults. - A national register-based cross-sectional study on economic, health system-related and patient-related characteristics

BACKGROUND

Potentially inappropriate medicines (PIM), i.e. medicines in which the potential harms may outweigh the benefits, use may be associated with e.g. hospitalization, outpatient visits and health care costs. As regional institutions are often responsible for financing pharmaceuticals, understanding the regional variation of PIM use could help to tackle the associated problems and costs.

OBJECTIVE

To explore regional variation in PIM use among older adults and the association with regional health-system related, patient-related and economic characteristics and the frequency of PIM use.

METHODS

This is a nation-wide study based on the Finnish Prescription Register. PIM use was defined according to the Finnish Meds75+ database and regional characteristics derived from national statistics.

RESULTS

Variation at the hospital district level was large, with the largest difference between the most and least PIM prescribing being 45.2%. The factors associated with high PIM prescribing were a higher share of women and a higher number of private doctor visits per inhabitant in a municipality. The factor associated to lower PIM prescribing was a higher share of Swedish-speaking population. The studied factors explained 23% of the municipal-level variation in PIM.

CONCLUSIONS

Large regional differences may lead to regional inequality in prescribing and in the distribution of pharmaceutical costs. As only a small share of the variation was explained by economic, health system-related and patient-related factors, the key reasons may lie in unobserved prescribing practices.

An Interpretable Algorithm on Post-injury Health Service Utilization Patterns to Predict Injury Outcomes

Purpose Post-injury health service utilization (HSU) contributes to injury outcomes, but limited studies investigated their relationship. This study aims to group injured patients in transport accidents based on minimal historical information of their HSU so that the groups are meaningfully associated with the outcome of interest. Methods The data include 20,692 injured patients who had compensation claims over 3 years. We propose a hybrid approach, combining unsupervised and supervised machine learning methods. Based on the first week post-injury data, we identify a proper clustering of patients best associated with total cost to recovery, as well as the discovery of HSU patterns. This allows developing models to accurately predict the outcome of interest using the discovered patterns. Furthermore, we propose to use decision tree classifiers to accurately classify future patients into the discovered clusters using their first week post-injury information. Results Our hybrid approach has identified eight patient groups. The compactness of the resulted clusters, assessed by Average Silhouette Width metric, is 0.71 indicating well-defined clusters. The resulted patient groups are highly predictive of injury outcomes. They improve the cost predictability more than twice in comparison with predictors such as gender, age and injury type. These groups also have substantial association with patients' recovery. The transparency and interpretability of decision trees allow integrating the resulting classification rules conveniently in operational processes. Conclusions This study provides a framework to discover knowledge and useful insights for health service providers and policy makers to control injury outcomes, and consequently to reduce the severity of transport accidents.

Characteristics of high-drug-cost beneficiaries of public drug plans in 9 Canadian provinces: a cross-sectional analysis

BACKGROUND

Drugs are the fastest growing cost in the Canadian health care system, owing to the increasing number of high-cost drugs. The objective of this study was to examine the characteristics of high-drug-cost beneficiaries of public drug plans across Canada relative to other beneficiaries.

METHODS

We conducted a cross-sectional study among public drug plan beneficiaries residing in all provinces except Quebec. We used the Canadian Institute for Health Information's National Prescription Drug Utilization Information System to identify all drugs dispensed to beneficiaries of public drug programs in 2016/17. We stratified the cohort into 2 groups: high-drug-cost beneficiaries (top 5% of beneficiaries based on annual costs) and other beneficiaries (remaining 95%). For each group, we reported total drug costs, prevalence of high-cost claims (> $1000), median number of drugs, proportion of beneficiaries aged 65 or more, the 10 most costly reimbursed medications and the 10 medications most commonly reimbursed. We reported estimates overall and by province.

RESULTS

High-drug-cost beneficiaries accounted for nearly half (46.5%) of annual spending, with an average annual spend of $14 610 per beneficiary, compared to $1570 among other beneficiaries. The median number of drugs dispensed was higher among high-drug-cost beneficiaries than among other beneficiaries (13 [interquartile range (IQR) 7-19] v. 8 [IQR 4-13]), and a much larger proportion of high-drug-cost beneficiaries than other beneficiaries received at least 1 high-cost claim (40.9% v. 0.6%). Long-term medications were the most commonly used medications for both groups, whereas biologics and antivirals were the most costly medications for high-drug-cost beneficiaries.

INTERPRETATION

High-drug-cost beneficiaries were characterized by the use of expensive medications and polypharmacy relative to other beneficiaries. Interventions and policies to help reduce spending need to consider both of these factors.

Improving Prediction of High-Cost Health Care Users with Medical Check-Up Data

Studies found that a small portion of the population spent the majority of health care resources, and they highlighted the importance of predicting high-cost users in the health care management and policy. Most prior research on high-cost user prediction models are based on diagnosis data with additional cost and health care utilization data to improve prediction accuracy. To further improve the prediction of high-cost users, researchers have been testing various new data sources such as self-reported health status data. In this study, we use three categories of medical check-up data, laboratory tests, self-reported medical history, and self-reported health behavior data to build high-cost user prediction models, and to assess the medical check-up features as predictors of high-cost users. Using three data-mining models, logistic regression, random forest, and neural network models, we show that under the diagnosis-based approach, medical check-up data marginally improve diagnosis-based prediction models. Under the cost-based approach, we find that medical check-up data improve cost-based prediction models marginally and medical check-up data can be a viable alternate data source to diagnosis data in predicting high-cost users.

Examining Patterns of Polypharmacy in Bipolar Disorder: Findings from the REAP-BD, Korea

Based on Korean data from the Research on Asian Psychotropic Prescription Pattern for Bipolar Disorder, this study tried to present prescription patterns in biopolar disorder (BD) and its associated clinical features. Based on the information obtained from the study with structured questions, the tendency of prescription pattern was studied and analyzed. Polypharmacy was predominant, including simple polypharmacy in 51.1% and complex polypharmacy in 34.2% of patients. Subjects associated with simple or complex polypharmacy were significantly younger, had higher inpatient settings, a larger portion of onset with manic episode, a shorter duration of untreated illness, a shorter duration of current episode, were more overweight, used less antidepressants and used more anxiolytics. These findings can suggest higher polypharmacy rate in more severe BD and highlight the necessity of monitoring the weight of subjects with polypharmacy.

Prevalence of care fragmentation among outpatients attending specialist clinics in a regional hospital in Singapore: a cross-sectional study

OBJECTIVE

To measure the extent of multispecialty care fragmentation among outpatients receiving specialist care and identify associated risk factors for fragmented care.

DESIGN

A retrospective cross-sectional study.

SETTING

Specialist outpatient clinics (SOCs) in a Singapore regional hospital.

PARTICIPANTS

A total of 40 333 patients aged 21 and above with at least two SOC visits in the year 2016. Data for 146 792 physician consultation visits were used in the analysis and visits for allied health services and medical procedures were excluded.

OUTCOME MEASURES

The Fragmentation of Care Index (FCI) was used to measure care fragmentation for specialist outpatients. Log-linear regression with stepwise selection was used to investigate the association between FCI and patient age, gender, race and Most Frequently Visited Specialty (MFVS), controlling for number of different specialities seen.

RESULTS

About 36% experienced fragmented care (FCI >0) and their mean FCI was 0.70 (SD=0.20). FCI was found to be positively associated with age (p<0.001). Patients who most frequently visited Haematology, Endocrinology and Anaesthesiology specialities were associated with more fragmented care while those who most frequently visited Medical Oncology, Ophthalmology and Orthopaedics Surgery specialities were associated with less fragmented care.

CONCLUSION

Multispecialty care fragmentation was found to be moderately high in the outpatient specialist clinics and was found to be associated with patients' age and certain medical specialties. With an ageing population and a rising prevalence of multimorbidity, healthcare providers should seek to eliminate unnecessary referrals to reduce the extent of care fragmentation.

Comparison of Canadian public medication insurance plans and the impact on out-of-pocket costs

BACKGROUND

Research from 2006 documented substantial variation in medication coverage for residents across Canada. Since then, some provinces have implemented major medication plan reforms. We aimed to update the information on publicly funded medication insurance plans available across Canada and to compare out-of-pocket costs across the country.

METHODS

We compared provincial medication insurance plans using data from public websites and other public source documents. Using 2 hypothetical clinical examples, we determined the amount and type of a patient's out-of-pocket costs for 5 different patient subtypes that varied based on medication burden, age and income.

RESULTS

Each province offers a plan to all residents. Cost-sharing is employed across all provinces. Some residents must pay a premium to receive insurance or must pay 100% of their medication costs until they reach a deductible amount, above which government funding covers a portion of medication costs. With the scenario of low medication burden (medication cost about $500), out-of-pocket costs ranged from $250 to $2100 for higher-income residents and from $0 to $700 for lower-income residents. With the scenario of high medication burden (medication cost about $1800), the corresponding ranges were $250-$2500 and $0-$1100. The variation was due to province of residence, age and income.

INTERPRETATION

Variations in out-of-pocket payments continue to exist across the provinces, with some groups facing high expenses. Further work is required to understand the impact of different cost-sharing mechanisms, develop policies to limit out-of-pocket expenses and improve provincial drug plans.