Predicting healthcare utilization using a pharmacy-based metric with the WHO's Anatomic Therapeutic Chemical algorithm

Prescribing pattern insights from a longitudinal study of older adult inpatients with polypharmacy and chronic non-cancer pain

BACKGROUND

The present study sought to determine the prevalence of chronic non-cancer pain (CNCP) among older adult inpatients with polypharmacy. It also aimed to analyse prescription patterns and assess the therapy adequacy and patient complexity for those with and without CNCP.

METHODS

This 4-year longitudinal study examined data from an exhaustive acute care hospital register on home-dwelling older adult patients (≥65) with polypharmacy. Commonly known combinations of potentially inappropriate medications were used to estimate therapy adequacy. Patient complexity was evaluated by comparing number of comorbidities and investigating physical and cognitive deficits.

RESULTS

We determined a prevalence of CNCP of 9.7% among all older adult inpatients with polypharmacy, rising to 11.3% for those aged ≥85. Overall, CNCP patients were prescribed more drugs and had more comorbidities and physical and cognitive deficits than patients without CNCP. Older adult patients with CNCP received more analgesics, greater quantities of opioids, paracetamol and co-analgesics and elevated opioid dosages. Older adult patients with CNCP aged ≥85 received fewer analgesics, opioids, non-steroidal anti-inflammatory drugs and co-analgesics but more paracetamol. Older adult patients with CNCP were prescribed more potentially inappropriate medications involving opioids. In particular, 24.5% received an opioid and a hypnotic (benzodiazepine or Z-drug), and 8.6% received an opioid and a gabapentinoid.

CONCLUSION

Observed differences in medication use between older adult inpatients with or without CNCP may be relevant for clinical practice. Potentially inadequate co-prescribing (such as hypnotics and opioids) affects a higher proportion of patients with CNCP and may have serious unintended consequences.

SIGNIFICANCE STATEMENT

This study describes differences in prescription patterns between people with and without chronic non-cancer pain in a large dataset of 20,422 discharges. The differences found may be relevant to clinical practice. In particular, high co-prescribing of opioids and hypnotics may have serious unintended consequences. Greater physical and cognitive deficits may indicate greater patient complexity, and appropriate interventions need to be developed to improve the management of this vulnerable patient group.

Development and internal validation of a prediction model for long-term opioid use-an analysis of insurance claims data

ABSTRACT

In the United States, a public-health crisis of opioid overuse has been observed, and in Europe, prescriptions of opioids are strongly increasing over time. The objective was to develop and validate a multivariable prognostic model to be used at the beginning of an opioid prescription episode, aiming to identify individual patients at high risk for long-term opioid use based on routinely collected data. Predictors including demographics, comorbid diseases, comedication, morphine dose at episode initiation, and prescription practice were collected. The primary outcome was long-term opioid use, defined as opioid use of either >90 days duration and ≥10 claims or >120 days, independent of the number of claims. Traditional generalized linear statistical regression models and machine learning approaches were applied. The area under the curve, calibration plots, and the scaled Brier score assessed model performance. More than four hundred thousand opioid episodes were included. The final risk prediction model had an area under the curve of 0.927 (95% confidence interval 0.924-0.931) in the validation set, and this model had a scaled Brier score of 48.5%. Using a threshold of 10% predicted probability to identify patients at high risk, the overall accuracy of this risk prediction model was 81.6% (95% confidence interval 81.2% to 82.0%). Our study demonstrated that long-term opioid use can be predicted at the initiation of an opioid prescription episode, with satisfactory accuracy using data routinely collected at a large health insurance company. Traditional statistical methods resulted in higher discriminative ability and similarly good calibration as compared with machine learning approaches.

Mortality among acute myocardial infarction patients admitted to hospitals on weekends as compared with weekdays in Taiwan

Weekend effect has been considered to be associated with poorer quality of care and patient's survival. For acute myocardial infarction (AMI) patients, the question of whether patients admitted during off-hours have worse outcomes as compared with patients admitted during on-hours is still inconclusive. We conducted this study to explore the weekend effect in AMI patients, using a nationwide insurance database in Taiwan. Using Taiwan National Health Insurance (NHI) claims database, we designed a retrospective cohort study, and extracted 184,769 incident cases of AMI through the NHI claims database between January 2006 and December 2014. We divided the patients into weekend admission group and weekday admission group. Patients were stratified as ST elevation/non-ST elevation AMI and receiving/not receiving percutaneous coronary intervention (PCI). We used a logistic regression model to examine the relative risk of in-hospital mortality and 1-year mortality which were obtained from the Taiwan National Death Registry between study groups. We found no difference between weekend group and weekday group for risk of in-hospital mortality (15.8% vs 16.2%, standardized difference 0.0118) and risk of 1-year mortality (30.2% vs 30.9%, standardized difference 0.0164). There was no statistically significant difference among all the comparisons through the multivariate logistic regression analysis adjusting for all the covariates and stratifying by the subtypes of AMI and whether or not executing PCI during hospitalization. As for AMI patients in Taiwan, admission on weekends or weekdays did not have a significant impact on either in-hospital mortality or 1-year cumulative mortality.

The framing of time-dependent machine learning models improves risk estimation among young individuals with acute coronary syndromes

Acute coronary syndrome (ACS) is a common cause of death in individuals older than 55 years. Although younger individuals are less frequently seen with ACS, this clinical event has increasing incidence trends, shows high recurrence rates and triggers considerable economic burden. Young individuals with ACS (yACS) are usually underrepresented and show idiosyncratic epidemiologic features compared to older subjects. These differences may justify why available risk prediction models usually penalize yACS with higher false positive rates compared to older subjects. We hypothesized that exploring temporal framing structures such as prediction time, observation windows and subgroup-specific prediction, could improve time-dependent prediction metrics. Among individuals who have experienced ACS (n_{global_cohort} = 6341 and n_yACS = 2242), the predictive accuracy for adverse clinical events was optimized by using specific rules for yACS and splitting short-term and long-term prediction windows, leading to the detection of 80% of events, compared to 69% by using a rule designed for the global cohort.

Economic Effects of Fixed-Dose Versus Loose-Dose Combination Therapy for Type 2 Diabetes Patients

OBJECTIVE

We examined the effects of fixe-dose combinations (FDCs) versus loose-dose combinations (LDCs) on costs from the payer and patient perspective and investigated potential channels contributing to differences in costs between the two modes of treatment.

METHODS

We investigated administrative data from 2017 to 2020 on diabetes patients in Germany. After using prospensity-score matching to remove dissimilarities between FDC and LDC patients, we compared changes in costs with a difference-in-differences approach. We analyzed pharmaceutical costs, inpatient and outpatient costs, other costs and total healthcare costs from the payer perspective, and co-payments from the patient perspective.

RESULTS

The sample comprised 1117 FDC and 1272 LDC patients. Regression analysis revealed that FDC therapy significantly increased antidiabetic pharmaceutical spending in the first year by 5.5% (p < 0.01), but decreased co-payments by 33% (p < 0.01) in the first and 44% (p < 0.01) in the second year. We also observed a trend towards higher outpatient spending in the first year. No significant differences were found with respect to inpatient or other costs. The increase in antidiabetic pharmaceutical spending did not contribute to a significant increase in total healthcare expenditure. We identified a shift of co-payments to the payer and higher adherence as possible mechanisms behind the increase in antidiabetic pharmaceutical spending.

CONCLUSION

Although FDC therapy increased disease-specific pharmaceutical spending in the short term, this increase did not lead to differences in total healthcare costs from the payer perspective. From the patient perspective, FDC therapy may be the preferred treatment approach, because of significant saving in co-payments, which is likely attributable to the elimination of one co-payment and therefore a shift in costs to the payer.

Using Data-Driven Approaches to Classify and Predict Health Care Spending in Patients With Gout Using Urate-Lowering Therapy

OBJECTIVE

Despite increasing overall health care spending over the past several decades, little is known about long-term patterns of spending among US patients with gout. Current approaches to assessing spending typically focus on composite measures or patients agnostic to disease state; in contrast, examining spending using longitudinal measures may better discriminate patients and target interventions to those in need. We used a data-driven approach to classify and predict spending patterns in patients with gout.

METHODS

Using insurance claims data from 2017-2019, we used group-based trajectory modeling to classify patients ages 40 years or older diagnosed with gout and treated with urate-lowering therapy (ULT) by their total health care spending over 2 years. We assessed the ability to predict membership in each spending group using logistic and generalized boosted regression with split-sample validation. Models were estimated using different sets of predictors and evaluated using C statistics.

RESULTS

In 57,980 patients, the mean ± SD age was 71.0 ± 10.5 years, and 17,194 patients (29.7%) were female. The best-fitting model included the following groups: minimal spending (13.2%), moderate spending (37.4%), and high spending (49.4%). The ability to predict groups was high overall (e.g., boosted C statistics with all predictors: minimal spending [0.89], moderate spending [0.78], and high spending [0.90]). Although average adherence was relatively high in the population, for the high-spending group, the most influential predictors were greater gout medication adherence and diabetes melllitus diagnosis.

CONCLUSION

We identified distinct long-term health care spending patterns in patients with gout using ULT with high accuracy. Several clinical predictors could be key areas for intervention, such as gout medication use or diabetes melllitus.

Core Concepts in Pharmacoepidemiology: Measures of drug utilization based on individual-level drug dispensing data

Background

Drug utilization studies are essential to facilitate rational drug use in the society.

Aim

In this review, we provide an overview of drug utilization measures that can be used with individual‐level drug dispensing data, referencing additional reading on the individual analysis. This is intended to serve as a primer for those new to drug utilization research and a shortlist from which researchers can identify useful analytical approaches when designing their drug utilization study.

Results and Discussion

We provide an overview of: (1) basic measures of drug utilization which are used to describe changes in drug use over time or compare drug use in different populations; (2) treatment adherence measures with specific focus on persistence and implementation; (3) how to measure drug combinations which is useful when assessing drug–drug interactions, concomitant treatment, and polypharmacy; (4) prescribing quality indicators and measures to assess variations in drug use which are useful tools to assess appropriate use of drugs; (5) proxies of prescription drug misuse and skewness in drug use; and (6) considerations when describing the characteristics of drug users or prescribers.

One-year follow-up healthcare costs of patients diagnosed with skin cancer in Germany: a claims data analysis

BACKGROUND

Routine skin cancer screening (SCS) is covered by the German statutory health insurance (SHI) since 2008. The objective of this study was to compare direct healthcare costs between patients in whom skin cancer was detected by routine SCS and patients in whom skin cancer was not detected by routine SCS.

METHODS

A retrospective observational study of administrative claims data from a large German SHI was performed. Patients with a diagnosis of malignant melanoma (MM) or non-melanoma skin cancer (NMSC) diagnosed in 2014 or 2015 were included. Costs were obtained for one year before and one year after diagnosis and analyzed in a difference-in-differences approach using regression models. Frequency matching was applied and risk adjustment was performed. Additional analyses were conducted, separately for specific age groups, excluding persons who died during the observation period and without taking costs for screening into consideration.

RESULTS

A total of 131,801 patients were included, of whom 13,633 (10.3%) had a diagnosis of MM and 118,168 (89.7%) had a diagnosis of NMSC. The description of total costs (without risk adjustment) shows lower mean total costs among patients whose skin cancer was detected via routine SCS compared to patients in whom skin cancer was not detected by routine SCS (MM: €5,326 (95% confidence interval (CI) €5,073; €5,579) vs. €9,038 (95% CI €8,629; €9,448); NMSC: €4,660 (95% CI €4,573; €4,745) vs. €5,890 (95% CI €5,813; €5,967)). Results of the regression analysis show cost savings of 18.8% (95% CI -23.1; -8.4) through routine SCS for patients with a diagnosis of MM. These cost savings in MM patients were more pronounced in patients younger than 65 years of age. For patients with a diagnosis of NMSC, the analysis yields a non-substantial increase in costs (2.5% (95% CI -0.1; 5.2)).

CONCLUSION

Cost savings were detected for persons with an MM diagnosed by routine SCS. However, the study could not detect lower costs due to routine SCS in the large fraction of persons with a diagnosis of NMSC. These results offer important insights into the cost structure of the routine SCS and provide opportunities for refinements.

Effectiveness of a home telemonitoring program for patients with chronic obstructive pulmonary disease in Germany: Evidence from the first three years

Introduction

Chronic obstructive pulmonary disease (COPD) affects more than 6 million people in Germany. Monitoring the vital parameters of COPD patients remotely through telemonitoring may help doctors and patients prevent and treat acute exacerbations of COPD, improving patients’ quality of life and saving costs for the statutory health insurance system.

Objective

To evaluate the effects from October 2012 until December 2015 of a structured home telemonitoring program implemented by a statutory health insurer in Germany.

Methods

We conducted a retrospective cohort study using administrative data. After building a balanced control group using Entropy Balancing, we calculated difference-in-difference estimators to account for time-invariant heterogeneity. We estimated differences in mortality rates using Cox regression and conducted subgroup and sensitivity analyses to check the robustness of the base case results. We observed each patient in the program for up to 3 years depending on his or her time of enrolment.

Results

Among patients in the telemonitoring cohort, we observed significantly higher inpatient costs due to COPD (€524.2, p<0,05; €434.6, p<0.05) and outpatient costs (102.5, p<0.01; 78.8 p<0.05) during the first two years of the program. Additional cost categories were significantly increased during the first year of telemonitoring. We also observed a significantly higher number of drug prescriptions during all three years of the observation period (2.0500, p < 0.05; 0.7260, p < 0.05; 3.3170, p < 0.01) and a higher number of outpatient contacts during the first two years (0.945, p<0.01, 0.683, p<0.05). Furthermore, we found significantly improved survival rates for participants in the telemonitoring program (HR 0.68, p<0.001).

Conclusion

On one hand, telemonitoring was associated with higher health care expenditures, especially in the first year of the program. For example, we were able to identify a statistically significant increase in inpatient costs due to COPD, outpatient contacts and drug prescriptions among individuals participating in the telemonitoring program. On the other hand, the telemonitoring program was accompanied by a survival benefit, which might be related to higher adherence rates, more intense treatment, or an improved understanding of COPD among these patients.

Predicting Future Service Use in Dutch Mental Healthcare: A Machine Learning Approach

A mental healthcare system in which the scarce resources are equitably and efficiently allocated, benefits from a predictive model about expected service use. The skewness in service use is a challenge for such models. In this study, we applied a machine learning approach to forecast expected service use, as a starting point for agreements between financiers and suppliers of mental healthcare. This study used administrative data from a large mental healthcare organization in the Netherlands. A training set was selected using records from 2017 (N = 10,911), and a test set was selected using records from 2018 (N = 10,201). A baseline model and three random forest models were created from different types of input data to predict (the remainder of) numeric individual treatment hours. A visual analysis was performed on the individual predictions. Patients consumed 62 h of mental healthcare on average in 2018. The model that best predicted service use had a mean error of 21 min at the insurance group level and an average absolute error of 28 h at the patient level. There was a systematic under prediction of service use for high service use patients. The application of machine learning techniques on mental healthcare data is useful for predicting expected service on group level. The results indicate that these models could support financiers and suppliers of healthcare in the planning and allocation of resources. Nevertheless, uncertainty in the prediction of high-cost patients remains a challenge.

Safety and survival of docetaxel and cabazitaxel in metastatic castration-resistant prostate cancer

OBJECTIVES

To investigate real-world haematological toxicity, overall survival (OS) and the treatment characteristics of docetaxel and cabazitaxel chemotherapy in metastatic castration-resistant prostate cancer (mCRPC).

PATIENTS AND METHODS

This retrospective claims data study followed patients with mCRPC receiving cabazitaxel or docetaxel from their first chemotherapy infusion. Haematological toxicities were measured using treatment codes and inpatient diagnoses. OS was estimated using the Kaplan-Meier method. A multivariable Cox regression analysis was used to identify OS predictors.

RESULTS

Data from 539 patients administered docetaxel and 240 administered cabazitaxel were analysed. Regarding adverse events, within 8 months of treatment initiation, some kind of treatment for haematological toxicity was documented in 31% of patients given docetaxel and in 61% of patients given cabazitaxel. In the same period, hospitalization associated with haematological toxicity was documented in 11% of the patients in the docetaxel cohort and in 15% of the patients in the cabazitaxel cohort. In the docetaxel cohort, 9.9% of patients required reverse isolation and 13% were diagnosed with sepsis during hospitalization. In the cabazitaxel cohort, the cumulative incidence was 7.9% and 15%, respectively. The median OS was reached at 21.9 months in the docetaxel cohort and, because of a later line of therapy, at 11.3 months in the cabazitaxel cohort. A multivariate Cox regression revealed that indicators of locally advanced and metastatic disease, severe comorbidities, and prior hormonal/cytotoxic therapies were independent predictors of early death.

CONCLUSION

Cabazitaxel patients face an increased risk of haematological toxicities during treatment. Together with their short survival time, this calls for a strict indication when using cabazitaxel in patients with mCRPC.

Treatment-Related Healthcare Costs of Metastatic Castration-Resistant Prostate Cancer in Germany: A Claims Data Study

PURPOSE

Treatments for patients with metastatic castration-resistant prostate cancer (mCRPC) have expanded rapidly. They include the chemotherapies docetaxel and cabazitaxel, hormonal drugs abiraterone and enzalutamide, and best supportive care (BSC). Cabazitaxel has proven to be the last life-prolonging option, associated with a significant risk of serious adverse events. Given the lack of real-world evidence, we aimed to compare healthcare resource utilization (HRU) and costs in patients with mCRPC treated with cabazitaxel, docetaxel, abiraterone, enzalutamide, and BSC.

METHODS

We used 2014-2017 claims data from a large German statutory health insurance fund, the Techniker Krankenkasse, to identify patients with mCRPC. Patient allocation to individual therapy regimens was based on clinical knowledge and included therapy cycles, duration of therapy, and continuous treatment. The study period lasted from the first claim until death, the end of data availability, a drug switch, or discontinuation of therapy, whichever came first. Multivariate regression models were used to compare monthly all-cause and mCRPC-related HRU and costs across cohorts by adjusting for baseline covariates (including age and comorbidities).

RESULTS

The 3944 identified patients with mCRPC initiated treatment with cabazitaxel (n = 240), docetaxel (n = 539), abiraterone (n = 486), enzalutamide (n = 351), or BSC (n = 2328). In most domains, HRU was highest in the cabazitaxel cohort and lowest in the BSC group. Accordingly, the highest all-cause and mCRPC-related costs per month, respectively, were observed in patients receiving cabazitaxel (€7631/€6343), followed by abiraterone (€5226/€4579), enzalutamide (€5079/€4416), docetaxel (€2392/€1580), and BSC (€959/€438). Cost variations were mostly attributable to drugs, inpatient treatment, and sick leave payments.

CONCLUSION

mCRPC treatment imposes a high economic burden on statutory health insurance. Cabazitaxel is associated with substantially higher expenses, resulting from higher drug costs and a greater need for inpatient treatment. As mCRPC continues to be incurable, decision makers and clinician leaders should carefully evaluate public access to innovative agents and optimal treatment strategies.

Mortality of major cardiovascular emergencies among patients admitted to hospitals on weekends as compared with weekdays in Taiwan

Background

Several studies have found a so-called weekend effect that patients admitted at the weekends had worse clinical outcomes than patients admitted at the weekdays. We performed this retrospective cohort study to explore the weekend effect in four major cardiovascular emergencies in Taiwan.

Methods

The Taiwan National Health Insurance (NHI) claims database between 2005 and 2015 was used. We extracted 3811 incident cases of ruptured aortic aneurysm, 184,769 incident cases of acute myocardial infarction, 492,127 incident cases of ischemic stroke, and 15,033 incident cases of pulmonary embolism from 9,529,049 patients having at least one record of hospitalization in the NHI claims database within 2006 ~ 2014. Patients were classified as weekends or weekdays admission groups. Dates of in-hospital mortality and one-year mortality were obtained from the Taiwan National Death Registry.

Results

We found no difference in in-hospital mortality between weekend group and weekday group in patients with ruptured aortic aneurysm (45.4% vs 45.3%, adjusted odds ratio [OR] 1.01, 95% confidence interval [CI] 0.87–1.17, p = 0.93), patients with acute myocardial infarction (15.8% vs 16.2%, adjusted OR 0.98, 95% CI 0.95–1.00, p = 0.10), patients with ischemic stroke (4.1% vs 4.2%, adjusted OR 0.99, 95% CI 0.96–1.03, p = 0.71), and patients with pulmonary embolism (14.6% vs 14.6%, adjusted OR 1.02, 95% CI 0.92–1.15, p = 0.66). The results remained for 1 year in all the four major cardiovascular emergencies.

Conclusions

We found no difference in either short-term or long-term mortality between patients admitted on weekends and patients admitted on weekdays in four major cardiovascular emergencies in Taiwan.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-021-06553-7.

Regimen simplification and medication adherence: Fixed-dose versus loose-dose combination therapy for type 2 diabetes

Background

Suboptimal patient adherence to pharmacological therapy of type 2 diabetes may be due in part to pill burden. One way to reduce pill burden in patients who need multiple medications is to use fixed-dose combinations. Our study aimed to compare the effects of fixed-dose combination versus loose-dose combination therapy on medication adherence and persistence, health care utilization, therapeutic safety, morbidities, and treatment modification in patients with type 2 diabetes over three years.

Methods

Using administrative data, we conducted a retrospective controlled cohort study comparing type 2 diabetes patients who switched from monotherapy to either a fixed-dose combination or a loose-dose combination. Adherence was assessed as the primary endpoint and calculated as the proportion of days covered with medication. After using entropy balancing to eliminate differences in observable baseline characteristics between the two groups, we applied difference-in-difference estimators for each outcome to account for time-invariant unobservable heterogeneity.

Results

Of the 990 type 2 diabetes patients included in our analysis, 756 were taking a fixed-dose combination and 234 were taking a loose-dose combination. We observed a statistically significantly higher change in adherence (year one: 0.22, p<0.001, year two: 0.25, p<0.001, and year three: 0.29, p<0.001) as well as higher persistence and a smaller change in the number of drug prescriptions in each of the three years in the fixed-dose combination group compared to the loose-dose combination group. The differences were most pronounced in patients who were poorly adherent, had a high pill burden, or did not have a severe concomitant disease.

Conclusion

Our results indicate that taking a fixed-dose combination can lead to a significant improvement in adherence to pharmacological therapy of type 2 diabetes compared to a loose-dose combination. In particular, these findings suggest that reducing pill burden may improve disease management among patients with more complex medication demand and patients who have demonstrated poor medication adherence.

More than two-dozen prescription drug-based risk scores are available for risk adjustment: A systematic review

OBJECTIVE

While several prescription drug-based risk indices have been developed, their design, performance, and application has not previously been synthesized.

STUDY DESIGN AND SETTING

We searched Ovid MEDLINE, CINAHL and Embase from inception through March 3, 2020 and included studies that developed or updated a prescription drug-based risk index. Two reviewers independently performed screening and extracted information on data source, study population, cohort sizes, outcomes, study methodology and performance. Predictive performance was evaluated using C statistics for binary outcomes and R² for continuous outcomes. The PROSPERO ID for this review is CRD42020165498.

RESULTS

Of 19,112 articles that were retrieved, 124 were full-text screened and 25 were included, each of which represented a de novo or updated drug-based index. The indices were customized to varied age groups and clinical populations and most commonly evaluated outcomes including mortality (36%), hospitalization (24%) and healthcare costs (24%). C statistics ranged from 0.62 to 0.92 for mortality and 0.59 to 0.72 for hospitalization, while adjusted R² for healthcare costs ranged from 0.06 to 0.62. Seven of the 25 risk indices included used global drug classification algorithms.

CONCLUSIONS

More than two-dozen prescription drug-based risk indices have been developed and they differ significantly in design, performance and application.

Costs of routine skin cancer screening in Germany: a claims data analysis

BACKGROUND

In 2008, a routine skin cancer screening (rSCS) programme was implemented in Germany. Since then, its medical and economical effects have been evaluated and critically discussed.

AIM

To compare costs for patients diagnosed with skin cancer with preceding rSCS vs. those diagnosed without rSCS.

METHODS

We conducted a retrospective observational study using claims data from a large German health insurance company for the period 2013-2016. We applied entropy balancing, difference-in-differences estimation and generalized linear models to compare costs for patients with cancer with and without rSCS. We conducted sensitivity analyses to test for the robustness of results.

RESULTS

In total, 12 790 patients with skin cancer were included in the analyses, of whom 6041 were diagnosed by rSCS. Treatment costs were €467 higher for patients in the control group (without rSCS). However, the screening costs were higher. For all people covered by the health insurance company, there were additional costs of €1339-1431 per patient with skin cancer diagnosed by rSCS. Thus, total costs, including costs for treatment and screening, were €872-964 higher for patients diagnosed by rSCS.

CONCLUSIONS

Based on our analysis, rSCS has the potential to reduce treatment costs; however, the screening costs exceed these savings.

Do discontinuities in marginal reimbursement affect inpatient psychiatric care in Germany?

This paper examines the behaviour of mental health care providers in response to marginal payment incentives induced by a discontinuous per diem reimbursement schedule with varying tariff rates over the length of stay. The analyses use administrative data on 12,627 cases treated in 82 psychiatric hospitals and wards in Germany. We investigate whether substantial reductions in marginal reimbursement per inpatient day led to strategic discharge behaviour once a certain length of stay threshold is exceeded. The data do not show gaps and bunches at the duration of treatment when marginal reimbursement decreases. Using logistic regression models, we find that providers did not react to discontinuities in marginal reimbursement by significantly reducing inpatient length of stay around the threshold. These findings are robust in terms of different model specifications and subsamples. The results indicate that if regulators aim to set incentives to decrease LOS, this might not be achieved by cuts in reimbursement over LOS.

Use of Data-Driven Methods to Predict Long-term Patterns of Health Care Spending for Medicare Patients

IMPORTANCE

Current approaches to predicting health care costs generally rely on a single composite value of spending and focus on short time horizons. By contrast, examining patients' spending patterns using dynamic measures applied over longer periods may better identify patients with different spending and help target interventions to those with the greatest need.

OBJECTIVE

To classify patients by their long-term, dynamic health care spending patterns using a data-driven approach and assess the ability to predict spending patterns, particularly using characteristics that are potentially modifiable through intervention.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study used a retrospective cohort design from a random nationwide sample of Medicare fee-for-service administrative claims data to identify beneficiaries aged 65 years or older with continuous eligibility from 2011 to 2013. Statistical analysis was performed from August 2018 to December 2019.

MAIN OUTCOMES AND MEASURES

Group-based trajectory modeling was applied to the claims data to classify the Medicare beneficiaries by their total health care spending patterns over a 2-year period. The ability to predict membership in each trajectory spending group was assessed using generalized boosted regression, a data mining approach to model building and prediction, with split-sample validation. Models were estimated using (1) prior-year predictors and (2) prior-year predictors potentially modifiable through intervention measured in the claims data. These models were evaluated using validated C-statistics. The relative influence of individual predictors in the models was evaluated.

RESULTS

Among the 329 476 beneficiaries, the mean (SD) age was 76.0 (7.2) years and 190 346 (57.8%) were female. This final 5-group model included a minimal-user group (group 1, 37 572 individuals [11.4%]), a low-cost group (group 2, 48 575 individuals [14.7%]), a rising-cost group (group 3, 24 736 individuals [7.5%]), a moderate-cost group (group 4, 83 338 individuals [25.3%]), and a high-cost group (group 5, 135 255 individuals [41.2%]). Potentially modifiable characteristics strongly predicted these patterns (C-statistics range: 0.68-0.94). For groups with progressively increasing spending in particular, the most influential factors were number of medications (relative influence: 29.2), number of office visits (relative influence: 30.3), and mean medication adherence (relative influence: 33.6).

CONCLUSIONS AND RELEVANCE

Using a data-driven approach, distinct spending patterns were identified with high accuracy. The potentially modifiable predictors of membership in the rising-cost group represent important levers for early interventions that may prevent later spending increases. This approach could be adapted by organizations to target quality improvement interventions, particularly because numerous health care organizations are increasingly using these routinely collected data.

Increased risk of adverse events in non-cancer patients with chronic and high-dose opioid use-A health insurance claims analysis

Background

Chronic and high dose opioid use may result in adverse events. We analyzed the risk associated with chronic and high dose opioid prescription in a Swiss population.

Methods

Using insurance claims data covering one-sixth of the Swiss population, we analyzed recurrent opioid prescriptions (≥2 opioid claims with at least 1 strong opioid claim) between 2006 and 2014. We calculated the cumulative dose in milligrams morphine equivalents (MED) and treatment duration. Excluded were single opioid claims, opioid use that was cancer treatment related, and opioid use in substitution programs. We assessed the association between the duration of opioid use, prescribed opioid dose, and benzodiazepine use with emergency department (ED) visits, urogenital and pulmonary infections, acute care hospitalization, and death at the end of the episode.

Results

In 63,642 recurrent opioid prescription episodes (acute 38%, subacute 7%, chronic 25.8%, very chronic (>360 days) episodes 29%) 18,336 ED visits, 30,209 infections, 19,375 hospitalizations, and 9,662 deaths occurred. The maximum daily MED dose was <20 mg in 15.8%, 20−<50 mg in 16.6%, 50−<100 mg in 21.6%, and ≥100 mg in 46%. Compared to acute episodes (<90 days), episode duration was an independent predictor of ED visits (chronic OR 1.09 (95% CI 1.03–1.15), very chronic (>360 days) OR 1.76 (1.67–1.86)) for adverse effects; infections (chronic OR 1.74 (1.66–1.82), very chronic 4.16 (3.95–4.37)), and hospitalization (chronic: OR 1.22 (1.16–1.29), very chronic OR 1.82 (1.73–1.93)). The risk of death decreased over time (very chronic OR 0.46 (0.43–0.50)). A dose dependent increased risk was observed for ED visits, hospitalization, and death (≥100mg daily MED OR 1.21 (1.13–1.29), OR 1.29 (1.21–1.38), and OR 1.67, 1.50–1.85, respectively). A concomitant use of benzodiazepines increased the odds for ED visits by 46% (OR 1.46, 1.41–1.52), infections by 44% (OR 1.44, 1.41–1.52), hospitalization by 12% (OR 1.12, 1.07–1.1), and death by 45% (OR 1.45, 1.37–1.53).

Conclusion

The length of opioid use and higher prescribed morphine equivalent dose were independently associated with an increased risk for ED visits and hospitalizations. The risk for infections, ED visits, hospitalizations, and death also increased with concomitant benzodiazepine use.

The economic burden of schizophrenia in Germany: A population-based retrospective cohort study using genetic matching

Objective

Prior studies to determine the economic consequences of schizophrenia have largely been undertaken in clinical settings with a small number of cases and have been unable to analyze effects across different age cohorts. The aim of this study is to investigate the burden of schizophrenia in Germany.

Methods

Costs, service utilization, and premature mortality attributable to schizophrenia were estimated for the year 2008 using a retrospective matched cohort design. Therefore, 26,977 control subjects as well as 9411 individuals with a confirmed diagnosis of schizophrenia were drawn from a sickness fund claims database. To reduce conditional bias, the non-parametric genetic matching method was employed.

Results

The final study population comprised 8224 matched pairs. The annual cost attributable to schizophrenia was €11,304 per patient from the payers’ perspective and €20,609 from the societal perspective with substantial variations among age groups: direct medical expenses were highest among patients aged > 65 years, whereas younger individuals (< 25 years) incurred the greatest non-medical costs. The annual burden of schizophrenia from the perspective of German society ranges between €9.63 billion and €13.52 billion.

Conclusion

There are considerable differences in the distribution of costs and service utilization for schizophrenia. Because schizophrenia is characterized by an early age of onset and a long duration, research efforts should be targeted at particular populations to obtain the most beneficial outcomes, both clinically and economically.

Prediction of health care expenditure increase: how does pharmacotherapy contribute?

BACKGROUND

Rising health care costs are a major public health issue. Thus, accurately predicting future costs and understanding which factors contribute to increases in health care expenditures are important. The objective of this project was to predict patients healthcare costs development in the subsequent year and to identify factors contributing to this prediction, with a particular focus on the role of pharmacotherapy.

METHODS

We used 2014-2015 Swiss health insurance claims data on 373'264 adult patients to classify individuals' changes in health care costs. We performed extensive feature generation and developed predictive models using logistic regression, boosted decision trees and neural networks. Based on the decision tree model, we performed a detailed feature importance analysis and subgroup analysis, with an emphasis on drug classes.

RESULTS

The boosted decision tree model achieved an overall accuracy of 67.6% and an area under the curve-score of 0.74; the neural network and logistic regression models performed 0.4 and 1.9% worse, respectively. Feature engineering played a key role in capturing temporal patterns in the data. The number of features was reduced from 747 to 36 with only a 0.5% loss in the accuracy. In addition to hospitalisation and outpatient physician visits, 6 drug classes and the mode of drug administration were among the most important features. Patient subgroups with a high probability of increase (up to 88%) and decrease (up to 92%) were identified.

CONCLUSIONS

Pharmacotherapy provides important information for predicting cost increases in the total population. Moreover, its relative importance increases in combination with other features, including health care utilisation.

The Economic Burden of Cystic Fibrosis in Germany from a Payer Perspective

BACKGROUND

Cystic fibrosis (CF) is a genetic disorder that is most common in white children and young adults. Long-term survival has improved steadily and will likely increase with the recent introduction of neonatal screening and causative treatment. However, these advances have substantial economic consequences for healthcare systems and payers.

OBJECTIVE

This study aims to determine the economic burden of CF and to elucidate the structure of costs and the distribution of resources for different subpopulations, treatment strategies and sites of care for CF in Germany.

METHODS

We conducted an observational cohort study to evaluate the economic burden of CF and the costs of treatment within different CF substrata from a payer perspective. Using claims data from a large German sickness fund, we identify the causal effect of CF on costs, service utilisation, and premature mortality. We compare the outcomes of a CF and a control group using entropy balancing and regression techniques, conduct further analyses for the CF group to gain insight into the economic burden associated with different levels of disease severity, and analyse pharmaceutical expenditures by collecting all CF-related expenses on outpatient drug treatment from the sickness fund database.

RESULTS

Direct medical costs caused by CF in Germany in 2016 average €17,551 per patient-year and appear to be mainly driven by the cost of outpatient drug prescriptions (€12,869). We estimate that the national burden of disease totals €159 million. Costs increase with disease severity and related complications. If all eligible CF patients in the German population were to receive CF mutation-specific drugs, the economic burden of disease would more than triple to €594 million.

CONCLUSION

CF has a constant and wide-ranging economic impact on payers, with considerable differences in the distribution of costs and service utilisation between younger and older patients as well as mild vs. severe patients. Pharmaceutical expenses will increase in the future as causative treatment gains importance. We encourage the use of a control group approach for burden-of-disease studies to be able to identify causal effects and thus to obtain more precise estimates.

Development of the "chronic condition measurement guide": a new tool to measure chronic conditions in older people based on ICD-10 and ATC-codes

PURPOSE

The aim of the study was to develop a comprehensive open-source measurement guide of the most prevalent chronic conditions among persons aged 65+ based on registry data of both diagnoses and prescribed drugs [the chronic condition measurement guide (CCMG)]. Furthermore, to investigate proof of concept of the measurement guide, different years of history and in- and excluding data on prescribed drugs. Finally, to investigate the measurement guide with other measurement guides designed to identify chronic conditions in persons aged 65+.

METHODS

The measurement guide was based on the 200 most prevalent chronic ICD10 codes in the Danish population 65+ years in 2015; the 200 most prevalent chronic ICD10 codes and causes of death in a cohort of 209,337 people who died of non-traumatic causes (January 2011-January 2016). Prescribed drugs were included in the measurement guide based on a literature review and specialist opinions.

RESULTS

We identified 83 different chronic conditions based on 1241 unique ICD-10 codes. Multimorbidity prevalence ranged from 10% (1-year history, excluding prescribing information) to 69% (15-year history, including prescribing information). We identified 95% of the persons with multimorbidity using the 29 most prevalent chronic conditions. Inclusion of these 29 conditions affected the prevalence of multimorbidity and 1-year mortality when the CCMG was compared with other measurement guides.

CONCLUSION

The CCMG is easily implemented using registry data. When implementing the measurement guide 10 years of history and drug prescribing information should be used. Using the CCMG to study multimorbidity, we recommend using at least the 29 most prevalent chronic conditions.

Healthcare cost prediction: Leveraging fine-grain temporal patterns

OBJECTIVE

To design and assess a method to leverage individuals' temporal data for predicting their healthcare cost. To achieve this goal, we first used patients' temporal data in their fine-grain form as opposed to coarse-grain form. Second, we devised novel spike detection features to extract temporal patterns that improve the performance of cost prediction. Third, we evaluated the effectiveness of different types of temporal features based on cost information, visit information and medical information for the prediction task.

MATERIALS AND METHODS

We used three years of medical and pharmacy claims data from 2013 to 2016 from a healthcare insurer, where the first two years were used to build the model to predict the costs in the third year. To prepare the data for modeling and prediction, the time series data of cost, visit and medical information were extracted in the form of fine-grain features (i.e., segmenting each time series into a sequence of consecutive windows and representing each window by various statistics such as sum). Then, temporal patterns of the time series were extracted and added to fine-grain features using a novel set of spike detection features (i.e., the fluctuation of data points). Gradient Boosting was applied on the final set of extracted features. Moreover, the contribution of each type of data (i.e., cost, visit and medical) was assessed. We benchmarked the proposed predictors against extant methods including those that used coarse-grain features which represent each time series with various statistics such as sum and the most recent portion of the values in the entire series. All prediction performances were measured in terms of Mean Absolute Percentage Error (MAPE).

RESULTS

Gradient Boosting applied on fine-grain predictors outperformed coarse-grain predictors with a MAPE of 3.02 versus 8.14 (p < 0.01). Enhancing the fine-grain features with the temporal pattern extraction features (i.e., spike detection features) further improved the MAPE to 2.04 (p < 0.01). Removing cost, visit and medical status data resulted in MAPEs of 10.24, 2.22 and 2.07 respectively (p < 0.01 for the first two comparisons and p = 0.63 for the third comparison).

CONCLUSIONS

Leveraging fine-grain temporal patterns for healthcare cost prediction significantly improves prediction performance. Enhancing fine-grain features with extraction of temporal cost and visit patterns significantly improved the performance. However, medical features did not have a significant effect on prediction performance. Gradient Boosting outperformed all other prediction models.

A real-world, population-based study of patterns of referral, treatment, and outcomes for advanced pancreatic cancer

Background

To describe patterns of referral, consultation, and treatment of advanced pancreatic cancer patients in a population‐based health care system and to evaluate the impact of these factors on outcomes.

Methods

This is a retrospective analysis of population‐based cancer data from the province of Alberta, Canada. We analyzed patients diagnosed with either locally advanced or metastatic pancreatic adenocarcinoma from 2009 to 2016 and evaluated their patterns of referral to a cancer center, consultation with oncology, and treatment with active anticancer therapies. Logistic regression models were constructed to determine the factors associated with referral, late oncology assessment, and late receipt of treatment.

Results

We identified 1621 pancreatic cancer patients. Median age was 70 years, 50% were men, and 51% had a Charlson index of 2+. Within this cohort, only 884 (54%) patients were referred to one of the provincial cancer centers. Adjusting for confounders in logistic regression models, older age and worse comorbidity scores were associated with nonreferral (both P < 0.01). In multivariable analysis among treated patients, the following factors were associated with improved overall survival, including younger age, earlier stage, and better comorbidity scores (all P < 0.01). Neither referral to consultation times nor consultation to treatment times correlated with outcomes. Importantly, nonreferred patients were more likely to use acute care services, including longer total duration of hospitalizations and more frequent visits with physician specialists.

Conclusion

A significant proportion of patients with advanced pancreatic cancer were never referred to a cancer center. Nonreferred patients were more likely to utilize specific health care resources.

Health Care Costs Associated With Incident Complications in Patients With Type 2 Diabetes in Germany

OBJECTIVE

The aim of this study is to provide reliable regression-based estimates of costs associated with different type 2 diabetes complications.

RESEARCH DESIGN AND METHODS

We used nationwide statutory health insurance (SHI) data from 316,220 patients with type 2 diabetes. Costs for inpatient and outpatient care, pharmaceuticals, rehabilitation, and nonmedical aids and appliances were assessed in the years 2013-2015. Quarterly observations are available for each year. We estimated costs (in 2015 euro) for complications using a generalized estimating equations model with a normal distribution adjusted for age, sex, occurrence of different complications, and history of complications at baseline, 2012. Two- and threefold interactions were included in an extended model.

RESULTS

The base case model estimated total costs in the quarter of event for the example of a 60- to 69-year-old man as follows: diabetic foot €1,293, amputation €14,284, retinopathy €671, blindness €2,933, nephropathy €3,353, end-stage renal disease (ESRD) €22,691, nonfatal stroke €9,769, fatal stroke €11,176, nonfatal myocardial infarction (MI)/cardiac arrest (CA) €8,035, fatal MI/CA €8,700, nonfatal ischemic heart disease (IHD) €6,548, fatal IHD €20,942, chronic heart failure €3,912, and angina pectoris €2,695. In the subsequent quarters, costs ranged from €681 for retinopathy to €6,130 for ESRD.

CONCLUSIONS

Type 2 diabetes complications have a significant impact on total health care costs in the SHI system, not only in the quarter of event but also in subsequent years. Men and women from different age-groups differ in their costs for complications. Our comprehensive estimates may support the parametrization of diabetes models and help clinicians and policy makers to quantify the economic burden of diabetes complications in the context of new prevention and treatment programs.

The impact of physician-level drug budgets on prescribing behavior

To contain pharmaceutical spending, drug budgets have been introduced across health systems. Apart from analyzing whether drug budgets fulfill their overall goal of reducing spending, changes in the cost and quality of prescribing and the enforcement mechanisms put in place need evaluation to assess the effectiveness of drug budgets at the physician level. In this study, we aim to analyze the cost and quality of prescribing conditional on the level of utilization of the drug budget and in view of varying levels of enforcement in cases of overspending. We observed drug budget utilization in a panel of 440 physicians in three federal states of Germany from 2005 to 2011. At the physician level, we retrospectively calculated drug budgets, the level of drug budget utilization, and differentiated by varying levels of enforcement where physicians overspent their budgets (i.e., more than 115/125% of the drug budget). Using lagged dependent-variable regression models, we analyzed whether the level of drug budget utilization in the previous year affected current prescribing in terms of various indicators to describe the cost and quality of prescribing. We controlled for patient and physician characteristics. The mean drug budget utilization is 92.3%. The level of drug budget utilization influences selected dimensions of cost and quality of prescribing (i.e., generic share (estimate 0.000215; p = 0.0246), concentration of generic brands (estimate 0.000585; p = 0.0056) and therapeutic substances (estimate -0.000060; p < 0.0001) and the share of potentially inappropriate medicines in the elderly (estimate 0.001; p < 0.0001)), whereas the level of enforcement does not. Physicians seem to gradually adjust their prescription patterns, especially in terms of generic substitution.

Health-economic evaluation of home telemonitoring for COPD in Germany: evidence from a large population-based cohort

INTRODUCTION

Telemonitoring for COPD has gained much attention thanks to its potential of reducing morbidity and mortality, healthcare utilisation and costs. However, its benefit with regard to clinical and economic outcomes remains to be clearly demonstrated.

OBJECTIVE

To analyse the effect of Europe's largest COPD telemonitoring pilot project on direct medical costs, health resource utilisation and mortality at 12 months.

METHODS

We evaluated a population-based cohort using administrative data. Difference-in-difference estimators were calculated to account for time-invariant unobservable heterogeneity after removing dissimilarities in observable characteristics between the telemonitoring and control group with a reweighting algorithm.

RESULTS

The analysis comprised 651 telemonitoring participants and 7047 individuals in the standard care group. The mortality hazards ratio was lower in the intervention arm (HR 0.51, 95 % CI 0.30-0.86). Telemonitoring cut total costs by 895 € (p < 0.05) compared to COPD standard care, mainly driven by savings in COPD-related hospitalisations in (very) severe COPD patients (-1056 €, p < 0.0001). Telemonitoring enrolees used healthcare (all-cause and COPD-related) less intensely with shorter hospital stays, fewer inpatient stays and smaller proportions of people with emergency department visits and hospitalisations (all p < 0.0001). Reductions in mortality, costs and healthcare utilisation were greater for (very) severe COPD cases.

CONCLUSION

This is the first German study to demonstrate that telemonitoring for COPD is a viable strategy to reduce mortality, healthcare costs and utilisation at 12 months. Contrary to widespread fear, reducing the intensity of care does not seem to impact unfavourably on health outcomes. The evidence offers strong support for introducing telemonitoring as a component of case management.

Appropriate selection for omalizumab treatment in patients with severe asthma?

Background: Omalizumab improves asthma control in patients with uncontrolled severe allergic asthma; however, appropriate patient selection is crucial. Information in this field is sparse.

Objective: We aimed to estimate whether potential omalizumab candidates were appropriately selected according to guidelines, and the clinical effect of omalizumab treatment over time.

Design: We performed a retrospective observational study on adult patients with asthma treated with omalizumab during 2006–2015 at the Department of Respiratory Medicine at Odense University Hospital (OUH), Denmark. Data were obtained from the Electronic Patient Journal of OUH and Odense Pharmaco-Epidemiological Database. Guideline criteria for omalizumab treatment were used to evaluate the appropriateness of omalizumab candidate selection, and the Asthma Control Test (ACT) to assess the clinical effects of omalizumab at weeks 16 and 52 from treatment initiation.

Results: During the observation period, 24 patients received omalizumab, but only 10 patients (42%) fulfilled criteria recommended by international guidelines. The main reasons for not fulfilling the criteria were inadequately reduced lung function, insufficient number of exacerbations, and asthma standard therapy below Global Initiative for Asthma (GINA) step 4–5. Seventeen and 11 patients completed treatment at weeks 16 and 52, with a statistically significant increase in ACT score of 5.1 points [95% confidence interval (CI) 3.1–7.2, p = 0.0001] and 7.7 points (95% CI 4.3–11.1, p = 0.0005), respectively.

Conclusion: Only 42% of the omalizumab-treated patients were appropriately selected according to current guidelines. Still, as omalizumab showed significant improvement in asthma control over time, it is important to keep this drug in mind as an add-on to asthma therapy in well-selected patients.

Longitudinal Patterns of Spending Enhance the Ability to Predict Costly Patients: A Novel Approach to Identify Patients for Cost Containment

BACKGROUND

With rising health spending, predicting costs is essential to identify patients for interventions. Many of the existing approaches have moderate predictive ability, which may result, in part, from not considering potentially meaningful changes in spending over time. Group-based trajectory modeling could be used to classify patients into dynamic long-term spending patterns.

OBJECTIVES

To classify patients by their spending patterns over a 1-year period and to assess the ability of models to predict patients in the highest spending trajectory and the top 5% of annual spending using prior-year predictors.

SUBJECTS

We identified all fully insured adult members enrolled in a large US nationwide insurer and used medical and prescription data from 2009 to 2011.

RESEARCH DESIGN

Group-based trajectory modeling was used to classify patients by their spending patterns over a 1-year period. We assessed the predictive ability of models that categorized patients in the top fifth percentile of annual spending and in the highest spending trajectory, using logistic regression and split-sample validation. Models were estimated using investigator-specified variables and a proprietary risk-adjustment method.

RESULTS

Among 998,651 patients, in the best-performing model, prediction was strong for patients in the highest trajectory group (C-statistic: 0.86; R: 0.47). The C-statistic of being in the top fifth percentile of spending in the best-performing model was 0.82 (R: 0.26). Approaches using nonproprietary investigator-specified methods performed almost as well as other risk-adjustment methods (C-statistic: 0.81 vs. 0.82).

CONCLUSIONS

Trajectory modeling may be a useful way to predict costly patients that could be implementable by payers to improve cost-containment efforts.

Health effects of parental deaths among adults in Norway: Purchases of prescription medicine before and after bereavement

We analyse effects of parental deaths on the health of women and men aged 18–59 in 2004–2008, indicated by purchases of prescription medicines. Register data covering the entire Norwegian population were used, and fixed-effects models were estimated to control for unobserved time-invariant individual factors. A parent's death seemed to have immediate adverse consequences in both main age groups considered (18–39, 40–59), although effects were lower in the older group. Some results suggested that this health disadvantage widened with increasing time since the parent's death. However, effects were weak: the annual number of different medicines purchased was only increased by 1–7% as a result of losing a parent. Death of a parent was associated with an immediate increase in purchases of medication for mental diseases, and there were indications of a physical response as well. As time since the parental death increased, there was a decline in the purchase of medication for mental diseases, but an opposite trend with respect to medication for other diseases. On the whole, maternal and paternal deaths had the same impact, and effects on daughters and sons were of the same magnitude.

•

A parent’s death seemed to have immediate weakly adverse health consequences.

•

Some results suggested that this health disadvantage widened with time since death.

•

The effects on physical and mental health seemed to be different.

•

Purchases of prescription medicine were used as indicators of health.

•

Unobserved time-invariant individual factors were controlled for.

Fertility history and use of antidepressant medication in late mid-life: a register-based analysis of Norwegian women and men

OBJECTIVES

Life course influences on later life depression may include parenting trajectories. We investigate associations between number and timing of births and use of antidepressant medication in late mid-life using data on the whole Norwegian population.

METHODS

We estimated logistic regression models to analyse variations in the purchase of antidepressants between 2004 and 2008 by timing of births and number of children among women and men aged 45-73, using Norwegian population register data. We controlled for age, education, marital and partnership status, and (in some models) family background shared among siblings.

RESULTS

Mothers and fathers of two or more children were generally less likely to purchase antidepressants than the childless. Mothers who started childbearing before age 22 were an exception, although according to sibling models they were not more likely to purchase antidepressants. All models showed that women who became mothers before age 26 and had only one child had higher odds of medication purchase than the childless. Older age at first birth was generally associated with lower risks of antidepressant purchase.

CONCLUSION

This analysis of high-quality data for a national population indicates that early motherhood, childlessness and low parity are associated with higher usage of antidepressants in late mid-life. Our data did not allow identification of mediating pathways, and we lacked information on early mental and physical health and some other potentially important confounders not shared between siblings. Furthermore purchase of antidepressants is not a perfect indicator of depression. Those concerns aside, the results suggest complex effects of fertility on depression that merit further investigation.

To what extent are psychiatrists aware of the comorbid somatic illnesses of their patients with serious mental illnesses? - a cross-sectional secondary data analysis

Background

Somatic comorbidities are a serious problem in patients with severe mental illnesses. These comorbidities often remain undiagnosed for a long time. In Germany, physicians are not allowed to access patients’ health insurance data and do not have routine access to documentation from other providers of health care. Against this background, the objective of this article was to investigate psychiatrists’ knowledge of relevant somatic comorbidities in their patients with severe mental illnesses.

Methods

Cross-sectional secondary data analysis was performed using primary data from a prospective study evaluating a model of integrated care of patients with serious mental illnesses. The primary data were linked with claims data from health insurers. Patients’ diagnoses were derived on the basis of the ICD-10 and the Anatomical Therapeutic Chemical (ATC) classification system. Diabetes, hypertension, coronary artery disease (CAD), hyperlipidaemia, glaucoma, osteoporosis, polyarthritis and chronic obstructive pulmonary disease (COPD) were selected for evaluation. We compared the number of diagnoses reported in the psychiatrists’ clinical report forms with those in the health insurance data.

Results

The study evaluated records from 1,195 patients with severe mental illnesses. The frequency of documentation of hypertension ranged from 21% in claims data to 4% in psychiatrists’ documentation, for COPD from 12 to 0%, respectively, and for diabetes from 7 to 2%, respectively. The percentage of diagnoses deduced from claims data but not documented by psychiatrists ranged from 68% for diabetes and 83% for hypertension, to 90% for CAD to 98% for COPD.

Conclusions

The majority of psychiatrists participating in the integrated care programme were insufficiently aware of the somatic comorbidities of their patients. We support allowing physicians to access patients’ entire medical records to increase their knowledge of patients’ medical histories and, consequently, to increase the safety and quality of care.

Clonal Dissemination of Pseudomonas aeruginosa Sequence Type 235 Isolates Carrying blaIMP-6 and Emergence of blaGES-24 and blaIMP-10 on Novel Genomic Islands PAGI-15 and -16 in South Korea

A total of 431 Pseudomonas aeruginosa clinical isolates were collected from 29 general hospitals in South Korea in 2015. Antimicrobial susceptibility was tested by the disk diffusion method, and MICs of carbapenems were determined by the agar dilution method. Carbapenemase genes were amplified by PCR and sequenced, and the structures of class 1 integrons surrounding the carbapenemase gene cassettes were analyzed by PCR mapping. Multilocus sequence typing (MLST) and pulsed-field gel electrophoresis (PFGE) were performed for strain typing. Whole-genome sequencing was carried out to analyze P. aeruginosa genomic islands (PAGIs) carrying the bla_IMP-6, bla_IMP-10, and bla_GES-24 genes. The rates of carbapenem-nonsusceptible and carbapenemase-producing P. aeruginosa isolates were 34.3% (148/431) and 9.5% (41/431), respectively. IMP-6 was the most prevalent carbapenemase type, followed by VIM-2, IMP-10, and GES-24. All carbapenemase genes were located on class 1 integrons of 6 different types on the chromosome. All isolates harboring carbapenemase genes exhibited genetic relatedness by PFGE (similarity > 80%); moreover, all isolates were identified as sequence type 235 (ST235), with the exception of two ST244 isolates by MLST. The bla_IMP-6, bla_IMP-10, and bla_GES-24 genes were found to be located on two novel PAGIs, designated PAGI-15 and PAGI-16. Our data support the clonal spread of an IMP-6-producing P. aeruginosa ST235 strain, and the emergence of IMP-10 and GES-24 demonstrates the diversification of carbapenemases in P. aeruginosa in Korea.

Costs and outcomes of the German disease management programme (DMP) for chronic obstructive pulmonary disease (COPD)-A large population-based cohort study

INTRODUCTION

To curb costs and improve health outcomes in chronic obstructive pulmonary disease (COPD), a nationwide disease management programme (DMP) was introduced in Germany in 2005. Yet, its effectiveness has not been comprehensively evaluated.

OBJECTIVE

To examine the effects of the German COPD DMP over three years on costs and health resource utilisation from the payer perspective, process quality, morbidity and mortality.

METHODS

A retrospective, population-based cohort study design is applied, using administrative data. After eliminating differences in observable characteristics between the DMP and the control group with entropy balancing, difference-in-difference estimators were computed to account for time-invariant unobservable heterogeneity.

RESULTS

215,104 individuals were included into the analysis of whom 25,269 were enrolled in the DMP. DMP patients had a reduced mortality hazard ratio (0.89, 95%CI: 0.84-0.94) but incurred excess costs of €553 per year. DMP enrolees reveal higher healthcare utilisation with larger shares of individuals being hospitalised (3.14%), consulting an outpatient clinic due to exacerbations (11.13%) and pharmaceutical prescriptions (2.78). However, average length of hospitalisation due to COPD fell by 0.49 days, adherence to medication guidelines as well as indicators for morbidity improved.

CONCLUSION

The German COPD DMP achieved significant improvements in mortality, morbidity and process quality, but at higher costs. Given the low ICER per life year gained, DMP COPD may constitute a cost-effective option to promote COPD population health.

Initiation of glucose-lowering treatment decreases international normalized ratio levels among users of vitamin K antagonists: a self-controlled register study

ESSENTIALS: It is not known if initiation of glucose-lowering drugs alters the efficacy of vitamin K antagonists (VKA). We examined if glucose-lowering drugs affected international normalized ratio (INR) in VKA-treated patients. Upon initiating glucose-lowering drugs, 51% of patients had INR values below the therapeutic window. Monitoring of INR levels should be intensified upon initiation of glucose-lowering drugs.

BACKGROUND

It is not known whether initiation of antidiabetic treatment affects the effect of vitamin K antagonists (VKAs). It was previously shown that metformin affects the effect of one VKA, phenprocoumon.

OBJECTIVES

The aim of this study was to determine if initiation of glucose-lowering treatment affects the international normalized ratio (INR) and dose requirements of the anticoagulant VKAs warfarin and phenprocoumon.

PATIENTS/METHODS

We performed a self-controlled retrospective register-based study. A total of 118 patients commencing glucose-lowering treatment while being treated with warfarin or phenprocoumon were included in the study. We compared INR, dose/INR and proportion of patients with at least one sub-therapeutic INR measurement before and after initiation of glucose-lowering treatment.

RESULTS

Initiation of glucose-lowering treatment caused mean INR to decrease from 2.5 to 2.2 (decrease of -0.3 [95% CI: -0.1; -0.5]) and led to more than half of the patients having at least one sub-therapeutic INR measurement. Six to 12 weeks later, the VKA dose/INR was increased by 11%, indicating a weakened effect of the VKA.

CONCLUSION

Initiation of glucose-lowering treatment reduces the anticoagulant effect of VKAs to an extent that is likely to be clinically relevant. This finding needs confirmation and mechanistic explanation.

The potential drug-drug interaction between proton pump inhibitors and warfarin

BACKGROUND

Proton pump inhibitors (PPIs) have been suggested to increase the effect of warfarin, and clinical guidelines recommend careful monitoring of international normalized ratio (INR) when initiating PPI among warfarin users. However, this drug-drug interaction is sparsely investigated in a clinical setting. The aim was to assess whether initiation of PPI treatment among users of warfarin leads to increased INR values.

METHODS

The study was an observational self-controlled study from 1998 to 2012 leveraging data on INR measurements on patients treated with warfarin from primary care and outpatient clinics and their use of prescription drugs. Data were analyzed in 2015. We assessed INR, warfarin dose, and dose/INR ratio before and after initiating PPI treatment using the paired student's t-test.

RESULTS

We identified 305 warfarin users initiating treatment with PPIs. The median age was 71 years (interquartile range 63-78 years), and 64% were men. The mean INR in the 70 days prior to PPI initiation was 2.6 (95%CI 2.5-2.8) and 2.6 (95%CI 2.5-2.7) in the period 1-3 weeks after PPI initiation (p = 0.67). Further, neither mean warfarin dose nor the dose/INR ratios were significantly different before and after PPI initiation. Sensitivity analyses revealed no differences among individual PPIs.

CONCLUSIONS

We found no evidence of a clinically meaningful drug-drug interaction between PPIs and warfarin in a Northern European patient population of unselected patients from an everyday outpatient and primary care clinical setting. Thus, we do not support the recommendation to "cautiously monitor" users of warfarin initiating PPI treatment.

Underuse of medication for circulatory disorders among unmarried women and men in Norway?

Background

It is well established that unmarried people have higher mortality from circulatory diseases and higher all-cause mortality than the married, and these marital status differences seem to be increasing. However, much remains to be known about the underlying mechanisms. Our objective was to examine marital status differences in the purchase of medication for circulatory diseases, and risk factors for them, which may indicate underuse of such medication by some marital status groups.

Methods

Using data from registers covering the entire Norwegian population, we analysed marital status differences in the purchase of medicine for eight circulatory disorders by people aged 50-79 in 2004-2008. These differences were compared with those in circulatory disease mortality during 2004-2007, considered as indicating probable differences in disease burden.

Results

The unmarried had 1.4-2.8 times higher mortality from the four types of circulatory diseases considered. However, the never-married in particular purchased less medicine for these diseases, or precursor risk factors of these diseases, primarily because of a low chance of making a first purchase. The picture was more mixed for the divorced and widowed. Both groups purchased less of some of these medicines than the married, but, especially in the case of the widowed, relatively more of other types of medicine. In contrast to the never-married, divorced and widowed people were as least as likely as the married to make a first purchase, but adherence rates thereafter, indicated by continuing purchases, were lower.

Conclusion

The most plausible interpretation of the findings is that compared with married people, especially the never-married more often have circulatory disorders that are undiagnosed or for which they for other reasons underuse medication. Inadequate use of these potentially very efficient medicines in such a large population group is a serious public health challenge which needs further investigation. It is possible that marital status differences in use of medicines for circulatory disorders combined with an increasing importance of these medicines have contributed to the widening marital status gap in mortality observed in several countries. This also requires further investigation.

Using clinical variables and drug prescription data to control for confounding in outcome comparisons between hospitals

Background

Hospital discharge records are an essential source of information when comparing health outcomes among hospitals; however, they contain limited information on acute clinical conditions. Doubts remain as to whether the addition of clinical and drug consumption information would improve the prediction of health outcomes and reduce confounding in inter-hospital comparisons. The objective of the study is to compare the performance of two multivariate risk adjustment models, with and without clinical data and drug prescription information, in terms of their capability to a) predict short-term outcome rates and b) compare hospitals’ risk-adjusted outcome rates using two risk-adjustment procedures.

Methods

Observational, retrospective study based on hospital data collected at the regional level.

Two cohorts of patients discharged in 2010 from hospitals located in the Lazio Region, Italy: acute myocardial infarction (AMI) and hip fracture (HF). Multivariate logistic regression models were implemented to predict 30-day mortality (AMI) or 48-hour surgery (HF), adjusting for demographic characteristics and comorbidities plus clinical data and drug prescription information. Risk-adjusted outcome rates were derived at the hospital level.

Results

The addition of clinical data and drug prescription information improved the capability of the models to predict the study outcomes for the two conditions investigated. The discriminatory power of the AMI model increases when the clinical data and drug prescription information are included (c-statistic increases from 0.761 to 0.797); for the HF model the increase was more slight (c-statistic increases from 0.555 to 0.574). Some differences were observed between the hospital-adjusted proportion estimated using the two different models. However, the estimated hospital outcome rates were weakly affected by the introduction of clinical data and drug prescription information.

Conclusions

The results show that the available clinical variables and drug prescription information were important complements to the hospital discharge data for characterising the acute severity of the patients. However, when these variables were used for adjustment purposes their contribution was negligible. This conclusion might not apply at other locations, in other time periods and for other health conditions if there is heterogeneity in the clinical conditions between hospitals.

Electronic supplementary material

The online version of this article (doi:10.1186/s12913-014-0495-3) contains supplementary material, which is available to authorized users.

Do short birth intervals have long-term implications for parental health? Results from analyses of complete cohort Norwegian register data

BACKGROUND

Short and very long interbirth intervals are associated with worse perinatal, infant and immediate maternal outcomes. Accumulated physiological, mental, social and economic stresses arising from raising children close in age may also mean that interbirth intervals have longer term implications for the health of mothers and fathers, but few previous studies have investigated this.

METHODS

Discrete-time hazards models were estimated to analyse associations between interbirth intervals and mortality risks for the period 1980-2008 in complete cohorts of Norwegian men and women born during 1935-1968 who had had two to four children. Associations between interbirth intervals and use of medication during 2004-2008 were also analysed using ordinary least-squares regression. Covariates included age, year, education, age at first birth, parity and change in coparent since the previous birth.

RESULTS

Mothers and fathers of two to three children with intervals between singleton births of less than 18 months, and mothers of twins, had raised mortality risks in midlife and early old age relative to parents with interbirth intervals of 30-41 months. For parents with three or four children, longer average interbirth intervals were associated with lower mortality. Short intervals between first and second births were also positively associated with medication use. Very long intervals were not associated with raised mortality or medication use when change of coparent since the previous birth was controlled.

CONCLUSIONS

Closely spaced and multiple births may have adverse long-term implications for parental health. Delayed entry to parenthood and increased use of fertility treatments mean that both are increasing, making this a public health issue which needs further investigation.

The economic impact of Marfan syndrome: a non-experimental, retrospective, population-based matched cohort study

Background

Marfan syndrome is a rare disease of the connective tissues, affecting multiple organ systems. Elevated morbidity and mortality in these patients raises the issue of costs for sickness funds and society. To date, there has been no study analysing the costs of Marfan syndrome from a sickness fund and societal perspective.

Objective

To estimate excess health resource utilisation, direct (non-)medical and indirect costs attributable to Marfan syndrome from a healthcare payer and a societal perspective in Germany in 2008.

Methods

A retrospective matched cohort study design is applied, using claims data. For isolating the causal effect of Marfan syndrome on excess costs, a genetic matching algorithm was used to reduce differences in observable characteristics between Marfan syndrome patients and the control group. 892 patients diagnosed with Marfan syndrome (ICD-10 Q87.4) were matched from a pool of 26,645 control individuals. After matching, we compared health resource utilisation and costs.

Results

From the sickness fund perspective, an average Marfan syndrome patient generates excess annual costs of €2496 compared with a control individual. From the societal perspective, excess annual costs amount to €15,728. For the sickness fund, the strongest cost drivers are inpatient treatment and care by non-physicians. From the sickness fund perspective, the third (25–41 years) and first (0–16 years) age quartiles reveal the greatest surplus in total costs. Marfan syndrome patients have 39% more physician contacts, a 153% longer average length of hospital stay, 119% more inpatient stays, 33% more prescriptions, 236% more medical imaging and 20% higher average prescription costs than control individuals. Depending on the prevalence, the economic impact from the sickness fund perspective ranges between €24.0 million and €61.4 million, whereas the societal economic impact extends from €151.3 million to €386.9 million.

Conclusions

Relative to its low frequency, Marfan syndrome requires high healthcare expenditure. Not only the high costs of Marfan syndrome but also its burden on patients’ lives call for more awareness from policy-makers, physicians and clinical researchers. Consequently, the diagnosis and treatment of Marfan syndrome should begin as soon as possible in order to prevent disease complications, early mortality and substantial healthcare expenditure.

Comparing potentially avoidable hospitalization rates related to ambulatory care sensitive conditions in Switzerland: the need to refine the definition of health conditions and to adjust for population health status

Background

Regional rates of hospitalization for ambulatory care sensitive conditions (ACSC) are used to compare the availability and quality of ambulatory care but the risk adjustment for population health status is often minimal. The objectives of the study was to examine the impact of more extensive risk adjustment on regional comparisons and to investigate the relationship between various area-level factors and the properly adjusted rates.

Methods

Our study is an observational study based on routine data of 2 million anonymous insured in 26 Swiss cantons followed over one or two years. A binomial negative regression was modeled with increasingly detailed information on health status (age and gender only, inpatient diagnoses, outpatient conditions inferred from dispensed drugs and frequency of physician visits). Hospitalizations for ACSC were identified from principal diagnoses detecting 19 conditions, with an updated list of ICD-10 diagnostic codes. Co-morbidities and surgical procedures were used as exclusion criteria to improve the specificity of the detection of potentially avoidable hospitalizations. The impact of the adjustment approaches was measured by changes in the standardized ratios calculated with and without other data besides age and gender.

Results

25% of cases identified by inpatient main diagnoses were removed by applying exclusion criteria. Cantonal ACSC hospitalizations rates varied from to 1.4 to 8.9 per 1,000 insured, per year. Morbidity inferred from diagnoses and drugs dramatically increased the predictive performance, the greatest effect found for conditions linked to an ACSC. More visits were associated with fewer PAH although very high users were at greater risk and subjects who had not consulted at negligible risk. By maximizing health status adjustment, two thirds of the cantons changed their adjusted ratio by more than 10 percent. Cantonal variations remained substantial but unexplained by supply or demand.

Conclusion

Additional adjustment for health status is required when using ACSC to monitor ambulatory care. Drug-inferred morbidities are a promising approach.

Disease identification based on ambulatory drugs dispensation and in-hospital ICD-10 diagnoses: a comparison

BACKGROUND

Pharmacy-based case mix measures are an alternative source of information to the relatively scarce outpatient diagnoses data. But most published tools use national drug nomenclatures and offer no head-to-head comparisons between drugs-related and diagnoses-based categories. The objective of the study was to test the accuracy of drugs-based morbidity groups derived from the World Health Organization Anatomical Therapeutic Chemical Classification of drugs by checking them against diagnoses-based groups.

METHODS

We compared drugs-based categories with their diagnoses-based analogues using anonymous data on 108,915 individuals insured with one of four companies. They were followed throughout 2005 and 2006 and hospitalized at least once during this period. The agreement between the two approaches was measured by weighted kappa coefficients. The reproducibility of the drugs-based morbidity measure over the 2 years was assessed for all enrollees.

RESULTS

Eighty percent used a drug associated with at least one of the 60 morbidity categories derived from drugs dispensation. After accounting for inpatient under-coding, fifteen conditions agreed sufficiently with their diagnoses-based counterparts to be considered alternative strategies to diagnoses. In addition, they exhibited good reproducibility and allowed prevalence estimates in accordance with national estimates. For 22 conditions, drugs-based information identified accurately a subset of the population defined by diagnoses.

CONCLUSIONS

Most categories provide insurers with health status information that could be exploited for healthcare expenditure prediction or ambulatory cost control, especially when ambulatory diagnoses are not available. However, due to insufficient concordance with their diagnoses-based analogues, their use for morbidity indicators is limited.

Identifying patients with chronic conditions using pharmacy data in Switzerland: an updated mapping approach to the classification of medications

Background

Quantifying population health is important for public health policy. Since national disease registers recording clinical diagnoses are often not available, pharmacy data were frequently used to identify chronic conditions (CCs) in populations. However, most approaches mapping prescribed drugs to CCs are outdated and unambiguous. The aim of this study was to provide an improved and updated mapping approach to the classification of medications. Furthermore, we aimed to give an overview of the proportions of patients with CCs in Switzerland using this new mapping approach.

Methods

The database included medical and pharmacy claims data (2011) from patients aged 18 years or older. Based on prescription drug data and using the Anatomical Therapeutic Chemical (ATC) classification system, patients with CCs were identified by a medical expert review. Proportions of patients with CCs were calculated by sex and age groups. We constructed multiple logistic regression models to assess the association between patient characteristics and having a CC, as well as between risk factors (diabetes, hyperlipidemia) for cardiovascular diseases (CVD) and CVD as one of the most prevalent CCs.

Results

A total of 22 CCs were identified. In 2011, 62% of the 932′612 subjects enrolled have been prescribed a drug for the treatment of at least one CC. Rheumatologic conditions, CVD and pain were the most frequent CCs. 29% of the persons had CVD, 10% both CVD and hyperlipidemia, 4% CVD and diabetes, and 2% suffered from all of the three conditions. The regression model showed that diabetes and hyperlipidemia were strongly associated with CVD.

Conclusions

Using pharmacy claims data, we developed an updated and improved approach for a feasible and efficient measure of patients’ chronic disease status. Pharmacy drug data may be a valuable source for measuring population’s burden of disease, when clinical data are missing. This approach may contribute to health policy debates about health services sources and risk adjustment modelling.

The influence of socio-economic status and multimorbidity patterns on healthcare costs: a six-year follow-up under a universal healthcare system

INTRODUCTION

Multimorbidity has been linked to elevated healthcare utilization and previous studies have found that socioeconomic status is an important factor associated with multimorbidity. Nonetheless, little is known regarding the impact of multimorbidity and socioeconomic status on healthcare costs and whether inequities in healthcare exist between socioeconomic classes within a universal healthcare system.

METHODS

This longitudinal study employed the claims database of the National Health Insurance of Taiwan (959 990 enrolees), adopting medication-based Rx-defined morbidity groups (Rx-MG) as a measurement of multimorbidity. Mixed linear models were used to estimate the effects of multimorbidity and socioeconomic characteristics on annual healthcare costs between 2005 and 2010.

RESULTS

The distribution of Rx-MGs and total costs presented statistically significant differences among gender, age groups, occupation, and income class (p < .001). Nearly 80% of the enrolees were classified as multimorbid and low income earners presented the highest prevalence of multimorbidity. After controlling for age and gender, increases in the number of Rx-MG assignments were associated with higher total healthcare costs. After controlling for the effects of Rx-MG assignment and demographic characteristics, physicians, paramedical personnel, and public servant were found to generate higher total costs than typical employees/self-employed enrolees, while low-income earners generated lower costs. High income levels were also found to be associated with lower total costs. It was also revealed that occupation and multimorbidity have a moderating effect on healthcare cost.

CONCLUSIONS

Increases in the prevalence of multimorbidity are associated with higher health care costs. This study determined that instances of multimorbidity varied according to socioeconomic class; likewise there were inequities in healthcare utilization among individuals of various occupations and income levels, even when demographic characteristics and multimorbidity were controlled for. This highlights the importance of socioeconomic status with regard to healthcare utilization. These results indicate that socioeconomic factors should not be discounted when discussing the utilization of healthcare by patients with multimorbidity.

The influence of socio-economic status and multimorbidity patterns on healthcare costs: a six-year follow-up under a universal healthcare system

Introduction

Multimorbidity has been linked to elevated healthcare utilization and previous studies have found that socioeconomic status is an important factor associated with multimorbidity. Nonetheless, little is known regarding the impact of multimorbidity and socioeconomic status on healthcare costs and whether inequities in healthcare exist between socioeconomic classes within a universal healthcare system.

Methods

This longitudinal study employed the claims database of the National Health Insurance of Taiwan (959 990 enrolees), adopting medication-based Rx-defined morbidity groups (Rx-MG) as a measurement of multimorbidity. Mixed linear models were used to estimate the effects of multimorbidity and socioeconomic characteristics on annual healthcare costs between 2005 and 2010.

Results

The distribution of Rx-MGs and total costs presented statistically significant differences among gender, age groups, occupation, and income class (p < .001). Nearly 80% of the enrolees were classified as multimorbid and low income earners presented the highest prevalence of multimorbidity. After controlling for age and gender, increases in the number of Rx-MG assignments were associated with higher total healthcare costs. After controlling for the effects of Rx-MG assignment and demographic characteristics, physicians, paramedical personnel, and public servant were found to generate higher total costs than typical employees/self-employed enrolees, while low-income earners generated lower costs. High income levels were also found to be associated with lower total costs. It was also revealed that occupation and multimorbidity have a moderating effect on healthcare cost.

Conclusions

Increases in the prevalence of multimorbidity are associated with higher health care costs. This study determined that instances of multimorbidity varied according to socioeconomic class; likewise there were inequities in healthcare utilization among individuals of various occupations and income levels, even when demographic characteristics and multimorbidity were controlled for. This highlights the importance of socioeconomic status with regard to healthcare utilization. These results indicate that socioeconomic factors should not be discounted when discussing the utilization of healthcare by patients with multimorbidity.