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

Harnessing the Potential of Real-World Evidence in the Treatment of Colorectal Cancer: Where Do We Stand?

OPINION STATEMENT

Treatment guidelines for colorectal cancer (CRC) are primarily based on the results of randomized clinical trials (RCTs), the gold standard methodology to evaluate safety and efficacy of oncological treatments. However, generalizability of trial results is often limited due to stringent eligibility criteria, underrepresentation of specific populations, and more heterogeneity in clinical practice. This may result in an efficacy-effectiveness gap and uncertainty regarding meaningful benefit versus treatment harm. Meanwhile, conduct of traditional RCTs has become increasingly challenging due to identification of a growing number of (small) molecular subtypes. These challenges-combined with the digitalization of health records-have led to growing interest in use of real-world data (RWD) to complement evidence from RCTs. RWD is used to evaluate epidemiological trends, quality of care, treatment effectiveness, long-term (rare) safety, and quality of life (QoL) measures. In addition, RWD is increasingly considered in decision-making by clinicians, regulators, and payers. In this narrative review, we elaborate on these applications in CRC, and provide illustrative examples. As long as the quality of RWD is safeguarded, ongoing developments, such as common data models, federated learning, and predictive modelling, will further unfold its potential. First, whenever possible, we recommend conducting pragmatic trials, such as registry-based RCTs, to optimize generalizability and answer clinical questions that are not addressed in registrational trials. Second, we argue that marketing approval should be conditional for patients who would have been ineligible for the registrational trial, awaiting planned (non) randomized evaluation of outcomes in the real world. Third, high-quality effectiveness results should be incorporated in treatment guidelines to aid in patient counseling. We believe that a coordinated effort from all stakeholders is essential to improve the quality of RWD, create a learning healthcare system with optimal use of trials and real-world evidence (RWE), and ultimately ensure personalized care for every CRC patient.

Treatment with psychostimulants and atomoxetine in people with psychotic disorders: reassessing the risk of clinical deterioration in a real-world setting

BACKGROUND

Although attention-deficit hyperactivity disorder (ADHD) is often comorbid with schizophrenia spectrum and other psychotic disorders (SZSPD), concerns about an increased risk of psychotic events have limited its treatment with either psychostimulants or atomoxetine.

AIMS

To examine whether the risk of hospital admission for psychosis in people with SZSPD was increased during the year following the introduction of such medications compared with the year before.

METHOD

This was a retrospective cohort study using Quebec (Canada) administrative health registries, including all Quebec residents with a public prescription drug insurance plan and a diagnosis of psychotic disorder, defined by relevant ICD-9 or ICD-10 codes, who initiated either methylphenidate, amphetamines or atomoxetine, between January 2010 and December 2016, in combination with antipsychotic medication. The primary outcome was time to hospital admission for psychosis within 1 year of initiation. State sequence analysis was also used to visualise admission trajectories for psychosis in the year following initiation of these medications, compared with the previous year.

RESULTS

Out of 2219 individuals, 1589 (71.6%) initiated methylphenidate, 339 (15.3%) amphetamines and 291 (13.1%) atomoxetine during the study period. After adjustment, the risk of hospital admission for psychosis was decreased during the 12 months following the introduction of these medications when used in combination with antipsychotics (adjusted HR = 0.36, 95% CI 0.24-0.54; P < 0.0001).

CONCLUSIONS

These findings suggest that, in a real-world setting, when used concurrently with antipsychotic medication, methylphenidate, amphetamines and atomoxetine may be safer than generally believed in individuals with psychotic disorders.

The Therapeutic Effect of Transcranial Magnetic Stimulation on Post-stroke Aphasia and the Optimal Treatment Parameters: A Meta-analysis

OBJECTIVE

This meta-analysis aimed to examine the overall effectiveness of TMS on post-stroke aphasia using a meta-analysis approach, as well as examine the effect of moderating variables (eg, study design, TMS protocol) on the effectiveness of TMS.

DATA SOURCES

A keyword search was conducted in 5 databases: ERIC, Google Scholar, PsycINFO, MEDLINE, and ProQuest (01/1985-12/2022). The search algorithm included all possible combinations of relevant keywords. Full-text articles were thoroughly examined using forward/backward search methods.

STUDY SELECTION

Studies were thoroughly screened using the following inclusion criteria: patients were diagnosed with post-stroke aphasia; studies focused on the effect of TMS on post-stroke aphasia; language assessments were conducted at pretest and posttest for TMS treatment and data were reported; studies included both an experimental group (ie, a group with TMS treatment) and a control group (ie, a group without TMS treatment).

DATA EXTRACTION

Information was extracted from each study including authors, publication year, first language of participants, study design, stroke duration, demographics of participants, TMS protocol, stimulation site, targeting, and statistical data of language performance pre- and post-TMS treatment.

DATA SYNTHESIS

A total of 17 studies were included in the final review, involving 682 patients with post-stroke aphasia (348 in the experimental group, 334 in the control group). The results showed that TMS treatment has significant immediate (Hedges' g=0.37) and maintenance (Hedges' g=0.34) effects on post-stroke aphasia. Additionally, the moderating variables showed a moderation effect on the effectiveness of TMS.

CONCLUSION

TMS treatment can significantly improve language ability for post-stroke aphasia. Additionally, this study provides an important reference for selecting the optimal TMS treatment parameters in treating post-stroke aphasia. Specifically, administering 15 sessions of rTMS lasting 10 min over the mirror area within Broca's area may produce the best TMS treatment outcomes.

Chronic Low Back Pain with and without Concomitant Osteoarthritis: A Retrospective, Longitudinal Cohort Study of Patients in England

Objective

Despite the high prevalence of chronic low back pain (CLBP) and osteoarthritis (OA), few estimates of the economic cost of these conditions in England have been published. The aim of the present analysis was to characterise the economic burden of moderate-to-severe pain associated with CLBP + OA and CLBP alone compared with general population-matched controls without CLBP or OA. The primary objective was to describe the total healthcare resource use (HCRU) and direct healthcare costs associated with the target patient populations. Secondary objectives were to describe treatment patterns and surgical procedures.

Methods

This was a retrospective, observational cohort study of patients receiving healthcare indicative of moderate-to-severe chronic pain associated with CLBP, with or without OA. We used linked longitudinal data from the Clinical Practice Research Datalink GOLD and Hospital Episode Statistics (HES). Patients (cases) were matched 1 : 1 with controls on age, sex, comorbidity burden, GP practice, and HES data availability.

Results

The CLBP-alone cohort comprised 13 554 cases with CLBP and 13 554 matched controls; the CLBP + OA cohort comprised 7803 cases with both OA and CLBP and 7803 matched controls. Across all follow-up periods, patients with CLBP alone and those with CLBP + OA had significantly more GP consultations, outpatient attendances, emergency department visits, and inpatient stays than controls (all p < 0.0001). By 36 months after indexing, the mean (SD) per-patient total direct healthcare cost in the CLBP-alone cohort was £5081 (£5905) for cases and £1809 (£4451) for controls (p < 0.0001); in the CLBP + OA cohort, the mean (SD) per-patient total direct healthcare cost was £8819 (£7143) for cases and £2428 (£4280) for controls (p < 0.0001).

Conclusion

Moderate-to-severe chronic pain associated with CLBP-with or without OA-has a substantial impact on patients and healthcare providers, leading to higher HCRU and costs versus controls among people with CLBP alone or together with OA.

Diagnostic properties of natriuretic peptides and opportunities for personalized thresholds for detecting heart failure in primary care

OBJECTIVES

Heart failure (HF) is a prevalent syndrome with considerable disease burden, healthcare utilization and costs. Timely diagnosis is essential to improve outcomes. This study aimed to compare the diagnostic performance of B-type natriuretic peptide (BNP) and N-terminal proBNP (NT-proBNP) in detecting HF in primary care. Our second aim was to explore if personalized thresholds (using age, sex, or other readily available parameters) would further improve diagnostic accuracy over universal thresholds.

METHODS

A retrospective study was performed among patients without prior HF who underwent natriuretic peptide (NP) testing in the Amsterdam General Practice Network between January 2011 and December 2021. HF incidence was based on registration out to 90 days after NP testing. Diagnostic accuracy was evaluated with AUROC, sensitivity and specificity based on guideline-recommended thresholds (125 ng/L for NT-proBNP and 35 ng/L for BNP). We used inverse probability of treatment weighting to adjust for confounding.

RESULTS

A total of 15,234 patients underwent NP testing, 6,870 with BNP (4.5 % had HF), and 8,364 with NT-proBNP (5.7 % had HF). NT-proBNP was more accurate than BNP, with an AUROC of 89.9 % (95 % CI: 88.4-91.2) vs. 85.9 % (95 % CI 83.5-88.2), with higher sensitivity (95.3 vs. 89.7 %) and specificity (59.1 vs. 58.0 %). Differentiating NP cut-off by clinical variables modestly improved diagnostic accuracy for BNP and NT-proBNP compared with a universal threshold.

CONCLUSIONS

NT-proBNP outperforms BNP for detecting HF in primary care. Personalized instead of universal diagnostic thresholds led to modest improvement.

Health economic impact of moderate-to-severe chronic pain associated with osteoarthritis in England: a retrospective analysis of linked primary and secondary care data

OBJECTIVE

Despite the prevalence of osteoarthritis (OA) in England, few studies have examined the health economic impact of chronic pain associated with OA. The aim of this study was to compare outcomes in patients with moderate-to-severe chronic pain associated with OA and matched controls without known OA.

DESIGN

Retrospective, longitudinal, observational cohort study.

SETTING

Electronic records extracted from the Clinical Practice Research Datalink GOLD primary care database linked to Hospital Episode Statistics (HES) data set.

PARTICIPANTS

Patients (cases; n=5931) ≥18 years and with existing diagnosis of OA and moderate-to-severe pain associated with their OA, and controls matched on age, sex, comorbidity burden, general practitioner (GP) practice and availability of HES data.

INTERVENTIONS

None.

PRIMARY AND SECONDARY OUTCOME MEASURES

Total healthcare resource use (HCRU) and direct healthcare costs during 0-6, 0-12, 0-24 and 0-36 months of follow-up. Secondary outcomes measures included pharmacological management and time to total joint replacement.

RESULTS

Patients with moderate-to-severe chronic pain associated with OA used significantly more healthcare services versus matched controls, reflected by higher HCRU and significantly higher direct costs. During the first 12 months' follow-up, cases had significantly more GP consultations, outpatient attendances, emergency department visits and inpatient stays than matched controls (all p<0.0001). Total mean costs incurred by cases during 0-12 months' follow-up were five times higher in cases versus controls (mean (SD): £4199 (£3966) vs £781 (£2073), respectively). Extensive cycling through pharmacological therapies was observed; among cases, 2040 (34.4%), 1340 (22.6%), 841 (14.2%), 459 (7.7%) and 706 (11.9%) received 1-5, 6-10, 11-15, 16-20 and >20 lines of therapy, respectively.

CONCLUSIONS

This wide-ranging, longitudinal, observational study of real-world primary and secondary care data demonstrates the impact of moderate-to-severe chronic pain associated with OA in patients compared with matched controls. Further studies are required to fully quantify the health economic burden of moderate-to-severe pain associated with OA.

A systematic review and meta-analysis of observational studies of the association between the use of incretin-based therapies and the risk of pancreatic cancer

BACKGROUND

Some early reports in the medical literature have raised concern about a possible increased risk of pancreatic cancer associated with the use of two broad classes of incretin-based therapies, dipeptidyl peptidase-4 inhibitors, and glucagon-like peptide-1 receptor agonists. This possibility has been somewhat mitigated by the null findings meta-analyses of randomized controlled trials, but the usefulness of their findings was hampered by serious shortcomings of lack of power and representativeness. These shortcomings can typically be addressed by observational studies, but observational studies on the topic have yielded conflicting findings. A systematic review and meta-analysis of observational studies was performed to qualitatively and quantitatively appraise the totality of evidence on the association between the use of incretin-based therapies and the risk of pancreatic cancer in routine clinical practice.

METHODS

The PubMed, Web of Science, Embase, and Google Scholar databases were searched. The study quality was appraised using the ROBINS-I tool and based on the presence of pharmacoepidemiology biases. A random-effects model was used to estimate the summary relative risks with corresponding CIs.

RESULTS

A total of 14 studies were included. The qualitative assessment revealed that all studies had inadequate follow-up (≤5 years), 12 studies were suspected to suffer from time-lag bias (due to inappropriate choice of comparator group) to varying extent, five studies included prevalent users, five studies did not implement exposure lag period, five studies had a serious risk of bias due to confounding, and one study had a time-window bias. The quantitative assessment showed no indication of an increased risk when all studies were pooled together (RR 1.04, 95% CI 0.87, 1.24) and when the analysis was restricted to the studies with the least bias (RR 0.77, 95% CI 0.51, 1.17). However, the pooled RRs were more frequently higher in the studies with less rigorous design and analysis. Specifically, a tendency toward an increased risk was observed in the studies with (RR 1.34, 95% CI 1.04, 1.72) or possibly with (RR 1.10, 95% CI 0.89, 1.36) time-lag bias, in the studies that did not apply (RR 1.23, 95% CI 0.93, 1.63) or with potentially inadequate exposure lag period of 6 months (RR 1.13, 95% CI 0.66, 1.94), in the studies that inappropriate comparator group of a combination of unspecified (RR 1.49, 95% CI 1.25, 1.78) or non-insulin (RR 1.15, 95% CI 0.93, 1.42) antidiabetic drugs, and in the studies with serious risk of bias due to confounding (RR 1.18, 95% CI 0.56, 2.49).

CONCLUSIONS

In summary, the totality of evidence from observational studies does not support the claim that the use of incretin-based therapies is associated with an increased risk of pancreatic cancer in routine clinical practice. The increased risk of pancreatic cancer observed in observational studies reflects bias resulting from suboptimal methodological approaches, which need to be avoided by future studies.

The Risk for Prostate Cancer With Calcium Channel Blockers: A Systematic Review, Meta-Analysis, and Meta-Regression

BACKGROUND

For decades, conflicting results were published regarding the increased risk of Prostate cancer (PCa) among calcium channel blocker (CCB) users.

OBJECTIVE

We aimed to evaluate the association between PCa and CCB exposure and assess moderating factors.

METHODS

We performed a systematic literature search in PubMed, Embase, and Cochrane databases for observational and randomized studies published until November 2020 with no language limitations, including data on the risk for PCa in CCB users compared with non-CCB users. We applied a random-effects model meta-analysis to pool results. In addition, we investigated potential moderating factors, such as CCB type, study type, participants' age, and duration of exposure, using meta-regression methods.

RESULTS

In our primary analysis, we included 18 studies. A statistically significant 5% increase in the risk for PCa was observed among CCB users (risk ratio [RR] = 1.05; 95% confidence interval [CI]: 1.01-1.10), with no significant association between the duration of exposure to CCBs and the risk for PCa (RR = 1.08; 95% CI: 0.98-1.19 for exposure for < 5years and RR = 1.01; 95% CI: 0.9-1.14 for exposure ≥ 5 years). The association remained statistically significant for the subgroup of dihydropyridines (RR = 1.13; 95% CI: 1.05-1.22). In addition, the association was not influenced by participants' age.

CONCLUSION AND RELEVANCE

CCBs are an important modality in treating hypertension. The 5% increased risk observed in the current meta-analysis could be influenced by residual confounding factors and should not affect hypertension treatment guidelines until more studies provide additional clinical information.

Patient-reported symptoms are a more reliable predictor of the societal burden compared to established physician-reported activity indices in inflammatory bowel disease: a cross-sectional study

BACKGROUND

The societal burden of inflammatory bowel diseases (IBD) is not well documented, and further studies are needed to quantify the costs of the disease state. Thus, the aim was to estimate the societal burden and identify its predictors.

METHODS

A cross-sectional questionnaire-based study complemented by objective data from patient medical records was performed for patients with Crohn's disease (CD) and ulcerative colitis (UC).

RESULTS

We analyzed data from 161 patients (CD: 102, UC: 59). The overall work impairment reached 15.4%, 11.2% vs. 28.8% without/with self-reported symptoms (p = 0.006). Daily activity impairment was 19.3%, 14.1% vs. 35.6% (p < 0.001). The disability pension rate was 28%, 23% vs. 44% (p = 0.012). The total productivity loss due to absenteeism, presenteeism, and disability amounted to 7,673 €/patient/year, 6,018 vs. 12,354 €/patient/year (p = 0.000). Out-of-pocket costs amounted to 562 €/patient/year, 472 vs. 844 €/patient/year (p = 0.001). Self-reported symptoms were the strongest predictor of costs (p < 0.001).

CONCLUSION

We found a high societal burden for IBD and a significant association between patient-reported disease symptoms and work disability, daily activity impairment, disability pensions, and out-of-pocket costs. Physician-reported disease activity is not a reliable predictor of costs except for out-of-pocket expenses.

Decision-analytic modeling for early health technology assessment of medical devices - a scoping review

Objective

The goal of this review was to identify decision-analytic modeling studies in early health technology assessments (HTA) of high-risk medical devices, published over the last three years, and to provide a systematic overview of model purposes and characteristics. Additionally, the aim was to describe recent developments in modeling techniques.

Methods

For this scoping review, we performed a systematic literature search in PubMed and Embase including studies published in English or German. The search code consisted of terms describing early health technology assessment and terms for decision-analytic models. In abstract and full-text screening, studies were excluded that were not modeling studies for a high-risk medical device or an in-vitro diagnostic test. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram was used to report on the search and exclusion of studies. For all included studies, study purpose, framework and model characteristics were extracted and reported in systematic evidence tables and a narrative summary.

Results

Out of 206 identified studies, 19 studies were included in the review. Studies were either conducted for hypothetical devices or for existing devices after they were already available on the market. No study extrapolated technical data from early development stages to estimate potential value of devices in development. All studies except one included cost as an outcome. Two studies were budget impact analyses. Most studies aimed at adoption and reimbursement decisions. The majority of studies were on in-vitro diagnostic tests for personalized and targeted medicine. A timed automata model, to our knowledge a model type new to HTA, was tested by one study. It describes the agents in a clinical pathway in separate models and, by allowing for interaction between the models, can reflect complex individual clinical pathways and dynamic system interactions. Not all sources of uncertainty for in-vitro tests were explicitly modeled. Elicitation of expert knowledge and judgement was used for substitution of missing empirical data. Analysis of uncertainty was the most valuable strength of decision-analytic models in early HTA, but no model applied sensitivity analysis to optimize the test positivity cutoff with regard to the benefit-harm balance or cost-effectiveness. Value-of-information analysis was rarely performed. No information was found on the use of causal inference methods for estimation of effect parameters from observational data.

Conclusion

Our review provides an overview of the purposes and model characteristics of nineteen recent early evaluation studies on medical devices. The review shows the growing importance of personalized interventions and confirms previously published recommendations for careful modeling of uncertainties surrounding diagnostic devices and for increased use of value-of-information analysis. Timed automata may be a model type worth exploring further in HTA. In addition, we recommend to extend the application of sensitivity analysis to optimize positivity criteria for in-vitro tests with regard to benefit-harm or cost-effectiveness. We emphasize the importance of causal inference methods when estimating effect parameters from observational data.

Causal evidence in health decision making: methodological approaches of causal inference and health decision science

Objectives

Public health decision making is a complex process based on thorough and comprehensive health technology assessments involving the comparison of different strategies, values and tradeoffs under uncertainty. This process must be based on best available evidence and plausible assumptions. Causal inference and health decision science are two methodological approaches providing information to help guide decision making in health care. Both approaches are quantitative methods that use statistical and modeling techniques and simplifying assumptions to mimic the complexity of the real world. We intend to review and lay out both disciplines with their aims, strengths and limitations based on a combination of textbook knowledge and expert experience.

Methods

To help understanding and differentiating the methodological approaches of causal inference and health decision science, we reviewed both methods with the focus on aims, research questions, methods, assumptions, limitations and challenges, and software. For each methodological approach, we established a group of four experts from our own working group to carefully review and summarize each method, followed by structured discussion rounds and written reviews, in which the experts from all disciplines including HTA and medicine were involved. The entire expert group discussed objectives, strengths and limitations of both methodological areas, and potential synergies. Finally, we derived recommendations for further research and provide a brief outlook on future trends.

Results

Causal inference methods aim for drawing causal conclusions from empirical data on the relationship of pre-specified interventions on a specific target outcome and apply a counterfactual framework and statistical techniques to derive causal effects of exposures or interventions from these data. Causal inference is based on a causal diagram, more specifically, a directed acyclic graph (DAG), which encodes the assumptions regarding the causal relations between variables. Depending on the type of confounding and selection bias, traditional statistical methods or more complex g-methods are needed to derive valid causal effects. Besides the correct specification of the DAG and the statistical model, assumptions such as consistency, positivity, and exchangeability must be checked when aiming at causal inference. Health decision science aims for guiding policy decision making regarding health interventions considering and balancing multiple competing objectives of a decision based on data from multiple sources and studies, for example prevalence studies, clinical trials and long-term observational routine effectiveness studies, and studies on preferences and costs. It involves decision analysis, a systematic, explicit and quantitative framework to guide decisions under uncertainty. Decision analyses are based on decision-analytic models to mimic the course of disease as well as aspects and consequences of the intervention in order to quantitatively optimize the decision. Depending on the type of decision problem, decision trees, state-transition models, discrete event simulation models, dynamic transmission models, or other model types are applied. Models must be validated against observed data, and comprehensive sensitivity analyses must be performed to assess uncertainty. Besides the appropriate choice of the model type and the valid specification of the model structure, it must be checked if input parameters of effects can be interpreted as causal parameters in the model. Otherwise results will be biased.

Conclusions

Both causal inference and health decision science aim for providing best causal evidence for informed health decision making. The strengths and limitations of both methods differ and a good understanding of both methods is essential for correct application but also for correct interpretation of findings from the described methods. Importantly, decision-analytic modeling should be combined with causal inference when developing guidance and recommendations regarding decisions on health care interventions.

Adoption and Trends in Uptake of Updated ICD-10 Codes for Clostridioides difficile-A Retrospective Observational Study

BACKGROUND

In October 2017, the single International Classification of Diseases, Tenth Revision (ICD-10), code for Clostridioides difficile infection (CDI), A04.7, was replaced with 2 codes delineating "recurrent CDI" (rCDI; A04.71) and "nonrecurrent CDI" (nrCDI; A04.72).

METHODS

To evaluate and validate use of the updated codes, this retrospective study included inpatient encounters with a CDI-related ICD-10 code from October 2016 to May 2019 in the PINC AITM Healthcare Database (PHD). Encounters after the October 2017 code update were characterized by clinical, facility, and provider variables and whether coding was concordant or discordant to the 8-week recurrence period. Multivariable regression analysis assessed variables associated with concordant coding.

RESULTS

Widespread adoption of the updated CDI codes across PHD hospitals occurred in October 2017. After October 2017, 21 446 CDI-related encounters met sample selection criteria (concordance in 67% of rCDI and 25% of nrCDI encounters). Higher proportions of rCDI- vs nrCDI-coded encounters (P < .05) had emergency room admission, admission by a gastroenterologist or infectious disease specialist, and were prescribed fidaxomicin, bezlotoxumab, or fecal microbiota transfer (FMT), with no significant difference by coding concordance status. Encounters coded concordantly were significantly more likely to be for rCDI (odds ratio [OR], 5.67; 95% CI, 5.32-6.03), a nonelective admission (OR, 1.35-1.69), or prescribed fidaxomicin (OR, 1.11; 95% CI, 1.01-1.23) or FMT (OR, 1.29; 95% CI, 1.17-1.42).

CONCLUSIONS

Our study findings suggest no delay in transition to the updated CDI-related codes. Treatment patterns for rCDI vs nrCDI encounters were consistent with Infectious Diseases Society of America guidelines, regardless of concordance status.

Real-world evidence analysis of the follicle-stimulating hormone use in male idiopathic infertility

Male idiopathic infertility remains a therapeutic challenge in the couple infertility management. In this setting, an empirical treatment with follicle-stimulating hormone (FSH) is allowed, although not recommended. Twenty-one clinical trials and four meta-analyses highlighted an overall increased pregnancy rate in case of FSH administration, but the indiscriminate FSH prescription is still unsupported by clinical evidence in idiopathic infertility. This context could represent an example in which real-world data (RWD) could add useful information. From a nationwide clinical practice survey performed in Italy, emerged the clinicians' attitude to prescribe FSH in the case of impaired semen with a significant improvement of semen parameters, identifying FSH treatment as a therapeutic card in the real-life management. Although more robust data are still needed to optimize FSH treatment in male idiopathic infertility, RWD should be included in the body of evidence considered in healthcare decision-making.

Nondifferential Treatment Misclassification Biases Toward the Null? Not a Safe Bet for Active Comparator Studies

Active comparator studies are increasingly common, particularly in pharmacoepidemiology. In such studies, the parameter of interest is a contrast (difference or ratio) in the outcome risks between the treatment of interest and the selected active comparator. While it may appear treatment is dichotomous, treatment is actually polytomous as there are at least 3 levels: no treatment, the treatment of interest, and the active comparator. Because misclassification may occur between any of these groups, independent nondifferential treatment misclassification may not be toward the null (as expected with a dichotomous treatment). In this work, we describe bias from independent nondifferential treatment misclassification in active comparator studies with a focus on misclassification that occurs between each active treatment and no treatment. We derive equations for bias in the estimated outcome risks, risk difference, and risk ratio, and we provide bias correction equations that produce unbiased estimates, in expectation. Using data obtained from US insurance claims data, we present a hypothetical comparative safety study of antibiotic treatment to illustrate factors that influence bias and provide an example probabilistic bias analysis using our derived bias correction equations.

HARmonized Protocol Template to Enhance Reproducibility of Hypothesis Evaluating Real-World Evidence Studies on Treatment Effects: A Good Practices Report of a Joint ISPE/ISPOR Task Force

OBJECTIVES

Ambiguity in communication of key study parameters limits the utility of real-world evidence (RWE) studies in healthcare decision-making. Clear communication about data provenance, design, analysis, and implementation is needed. This would facilitate reproducibility, replication in independent data, and assessment of potential sources of bias.

METHODS

The International Society for Pharmacoepidemiology (ISPE) and ISPOR-The Professional Society for Health Economics and Outcomes Research (ISPOR) convened a joint task force, including representation from key international stakeholders, to create a harmonized protocol template for RWE studies that evaluate a treatment effect and are intended to inform decision-making. The template builds on existing efforts to improve transparency and incorporates recent insights regarding the level of detail needed to enable RWE study reproducibility. The over-arching principle was to reach for sufficient clarity regarding data, design, analysis, and implementation to achieve 3 main goals. One, to help investigators thoroughly consider, then document their choices and rationale for key study parameters that define the causal question (e.g., target estimand), two, to facilitate decision-making by enabling reviewers to readily assess potential for biases related to these choices, and three, to facilitate reproducibility.

STRATEGIES TO DISSEMINATE AND FACILITATE USE

Recognizing that the impact of this harmonized template relies on uptake, we have outlined a plan to introduce and pilot the template with key international stakeholders over the next 2 years.

CONCLUSION

The HARmonized Protocol Template to Enhance Reproducibility (HARPER) helps to create a shared understanding of intended scientific decisions through a common text, tabular and visual structure. The template provides a set of core recommendations for clear and reproducible RWE study protocols and is intended to be used as a backbone throughout the research process from developing a valid study protocol, to registration, through implementation and reporting on those implementation decisions.

Methods and Study Design for Cancer Health Economics Research: Summary of Discussions From a Breakout Session

The legitimacy of findings from cancer health economics research depends on study design and methods. A breakout session, Methods and Study Design for Cancer Health Economics Research, was convened at the Future of Cancer Health Economics Research Conference to discuss 2 commonly used analytic tools for cancer health economics research: observational studies and decision-analytic modeling. Observational studies include analysis of data collected with the primary purpose of supporting economic evaluation or secondary use of data collected for another purpose. Modeling studies develop a parametrized structure, such as a decision tree, to estimate hypothetical impact. Whereas observational studies focus on what has happened and why, modeling studies address what may happen. We summarize the discussion at this breakout session, focusing on 3 key elements of high-quality cancer health economics research: study design, analytical methods, and addressing uncertainty.

Target Trial Emulation for Transparent and Robust Estimation of Treatment Effects for Health Technology Assessment Using Real-World Data: Opportunities and Challenges

Evidence about the relative effects of new treatments is typically collected in randomised controlled trials (RCTs). In many instances, evidence from RCTs falls short of the needs of health technology assessment (HTA). For example, RCTs may not be able to capture longer-term treatment effects, or include all relevant comparators and outcomes required for HTA purposes. Information routinely collected about patients and the care they receive have been increasingly used to complement RCT evidence on treatment effects. However, such routine (or real-world) data are not collected for research purposes, so investigators have little control over the way patients are selected into the study or allocated to the different treatment groups, introducing biases for example due to selection or confounding. A promising approach to minimise common biases in non-randomised studies that use real-world data (RWD) is to apply design principles from RCTs. This approach, known as 'target trial emulation' (TTE), involves (1) developing the protocol with respect to core study design and analysis components of the hypothetical RCT that would answer the question of interest, and (2) applying this protocol to the RWD so that it mimics the data that would have been gathered for the RCT. By making the 'target trial' explicit, TTE helps avoid common design flaws and methodological pitfalls in the analysis of non-randomised studies, keeping each step transparent and accessible. It provides a coherent framework that embeds existing analytical methods to minimise confounding and helps identify potential limitations of RWD and the extent to which these affect the HTA decision. This paper provides a broad overview of TTE and discusses the opportunities and challenges of using this approach in HTA. We describe the basic principles of trial emulation, outline some areas where TTE using RWD can help complement RCT evidence in HTA, identify potential barriers to its adoption in the HTA setting and highlight some priorities for future work.

Personalized Reimbursement Model (PRM) program: A real-world data platform of cancer drugs use to improve and personalize drug pricing and reimbursement in France

Objective

This article describes the Personalized Reimbursement Model (PRM) program methodology, limitations, achievement and perspectives in using real-world data of cancer drugs use to improve and personalize drug pricing and reimbursement in France.

Materials and methods

PRM platform aggregates Electronic Pharmacy Records (EPR) data from French medical centers (PRM centers) to build retrospective cohorts of patients treated with injectable cancer drugs in a hospital setting. Data extracted on January 1^st, 2020, from breast cancer (BC) patients who received trastuzumab, trastuzumab emtansin or pertuzumab since January 1st, 2011, and from lung cancer (LC) patients who received bevacizumab or atezolizumab since January 1^st, 2015, enabled recovering their injectable cancer drugs history from diagnosis date until December 30^th, 2019, and served as dataset for assessment.

Results

123 PRM centers provided data from 30,730 patients (25,660 BC and 5,070 LC patients respectively). Overall, 20,942 (82%) of BC and 4,716 (93%) of LC patients were analyzed. Completion rate was above 98% for patients characteristics, diagnostic and treatment related data. PRM centers cover 48% and 33% of BC and LC patients in-hospital therapeutic management in France, respectively. Distribution of BC and LC patients therapeutic management, by medical center category and geographic location, was similar in PRM centers to all French medical centers, ensuring the representativeness of the PRM platform.

Conclusion

PRM Platform enabled building a national database generating on demand Real-World Evidence based on EPR. This enabled the first performance-based risk-sharing arrangements based on PRM data, between the CEPS and Roche, for atezolizumab cancer immunotherapy in metastatic non-small cell lung cancer indication.

Framework for the synthesis of non-randomised studies and randomised controlled trials: a guidance on conducting a systematic review and meta-analysis for healthcare decision making

INTRODUCTION

High-quality randomised controlled trials (RCTs) provide the most reliable evidence on the comparative efficacy of new medicines. However, non-randomised studies (NRS) are increasingly recognised as a source of insights into the real-world performance of novel therapeutic products, particularly when traditional RCTs are impractical or lack generalisability. This means there is a growing need for synthesising evidence from RCTs and NRS in healthcare decision making, particularly given recent developments such as innovative study designs, digital technologies and linked databases across countries. Crucially, however, no formal framework exists to guide the integration of these data types.

OBJECTIVES AND METHODS

To address this gap, we used a mixed methods approach (review of existing guidance, methodological papers, Delphi survey) to develop guidance for researchers and healthcare decision-makers on when and how to best combine evidence from NRS and RCTs to improve transparency and build confidence in the resulting summary effect estimates.

RESULTS

Our framework comprises seven steps on guiding the integration and interpretation of evidence from NRS and RCTs and we offer recommendations on the most appropriate statistical approaches based on three main analytical scenarios in healthcare decision making (specifically, 'high-bar evidence' when RCTs are the preferred source of evidence, 'medium,' and 'low' when NRS is the main source of inference).

CONCLUSION

Our framework augments existing guidance on assessing the quality of NRS and their compatibility with RCTs for evidence synthesis, while also highlighting potential challenges in implementing it. This manuscript received endorsement from the International Society for Pharmacoepidemiology.

A methodological framework for tackling confounding by indication when assessing the treatment effects of Chinese herbal injections in the real world

AIM

In the context of integrative medicine, whether Chinese herbal injections are effective in routine practice has become a question of broad interest. However, confounding by indication (i.e., indication bias) is a prevalent and highly challenging methodological issue when using routinely collected health care data to assess the real-world effectiveness of Chinese herbal injections.

METHODS AND RESULTS

We proposed a methodological approach to tackling confounding by indication in assessing the real-world effectiveness of Chinese herbal injections, incorporating empirical experiences, a literature review and interactive discussions, and a panel of external experts to finally achieve a consensus. This approach consisted of three cohesive steps, including a full understanding of treatment patterns, construction of fair comparisons by identifying appropriate combination treatments and comparators, and using statistical methods to further control for confounding. In the investigation of treatment patterns, we proposed five domains to identify treatment patterns with Chinese herbal injections, and we offered five patterns of combination treatments to characterize how Chinese herbal injections are used in conjunction with other treatments. In constructing fair comparisons, we suggested the use of both nonuse and active comparators; given the diverse combination treatments, we developed six scenarios that may form fair comparisons. In the statistical analysis, we discussed five statistical models for controlling confounding by indication, including their pros and cons. We also included a practical example to illustrate the usefulness of the methodological approach.

CONCLUSION

The proposed approach may serve as an effective tool to guide researchers to reliably assess the effectiveness of Chinese herbal injections in the context of integrative medicine.

Reporting of abstracts in studies that used routinely collected data for exploring drug treatment effects: a cross-sectional survey

Background

In recent years, studies that used routinely collected data (RCD), such as electronic medical records and administrative claims, for exploring drug treatment effects, including effectiveness and safety, have been increasingly published. Abstracts of such studies represent a highly attended source for busy clinicians or policy-makers, and are important for indexing by literature database. If less clearly presented, they may mislead decisions or indexing. We thus conducted a cross-sectional survey to systematically examine how the abstracts of such studies were reported.

Methods

We searched PubMed to identify all observational studies published in 2018 that used RCD for assessing drug treatment effects. Teams of methods-trained collected data from eligible studies using pilot-tested, standardized forms that were developed and expanded from “The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology” (RECORD-PE) statement. We used descriptive analyses to examine how authors reported data source, study design, data analysis, and interpretation of findings.

Results

A total of 222 studies were included, of which 118 (53.2%) reported type of database used, 17 (7.7%) clearly reported database linkage, and 140 (63.1%) reported coverage of data source. Only 44 (19.8%) studies stated a predefined hypothesis, 127 (57.2%) reported study design, 140 (63.1%) reported statistical models used, 142 (77.6%) reported adjusted estimates, 33 (14.9%) mentioned sensitivity analyses, and 39 (17.6%) made a strong claim about treatment effect. Studies published in top 5 general medicine journals were more likely to report the name of data source (94.7% vs. 67.0%) and study design (100% vs. 53.2%) than those in other journals.

Conclusions

The under-reporting of key methodological features in abstracts of RCD studies was common, which would substantially compromise the indexing of this type of literature and prevent the effective use of study findings. Substantial efforts to improve the reporting of abstracts in these studies are highly warranted.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01482-9.

Effects of sustained weight loss on outcomes associated with obesity comorbidities and healthcare resource utilization

Objective

Determine the impact of long-term non-surgical weight loss maintenance on clinical relevance for osteoarthritis, cancer, opioid use, and depression/anxiety and healthcare resource utilization.

Methods

A cohort of adults receiving primary care within Geisinger Health System between 2001–2017 was retrospectively studied. Patients with ≥3 weight measurements in the two-year index period and obesity at baseline (BMI ≥30 kg/m²) were categorized: Obesity Maintainers (reference group) maintained weight within +/-3%; Weight Loss Rebounders lost ≥5% body weight in year one, regaining ≥20% of weight loss in year two; Weight Loss Maintainers lost ≥5% body weight in year one, maintaining ≥80% of weight loss. Association with development of osteoarthritis, cancer, opioid use, and depression/anxiety, was assessed; healthcare resource utilization was quantified. Magnitude of weight loss among maintainers was evaluated for impact on health outcomes.

Results

In total, 63,567 patients were analyzed including 67% Obesity Maintainers, 19% Weight Loss Rebounders, and 14% Weight Loss Maintainers; median follow-up was 9.7 years. Time until osteoarthritis onset was delayed for Weight Loss Maintainers compared to Obesity Maintainers (Logrank test p <0.0001). Female Weight Loss Maintainers had a 19% and 24% lower risk of developing any cancer (p = 0.0022) or obesity-related cancer (p = 0.0021), respectively. No significant trends were observed for opioid use. Weight loss Rebounders and Maintainers had increased risk (14% and 25%) of future treatment for anxiety/depression (both <0.0001). Weight loss maintenance of >15% weight loss was associated with the greatest decrease in incident osteoarthritis. Healthcare resource utilization was significantly higher for Weight Loss Rebounders and Maintainers compared to Obesity Maintainers. Increased weight loss among Weight Loss Maintainers trended with lower overall healthcare resource utilization, except for hospitalizations.

Conclusions

In people with obesity, sustained weight loss was associated with greater clinical benefits than regained short-term weight loss and obesity maintenance. Higher weight loss magnitudes were associated with delayed onset of osteoarthritis and led to decreased healthcare utilization.

The Structured Process to Identify Fit-for-purpose Data (SPIFD): A data feasibility assessment framework

To complement real‐world evidence (RWE) guidelines, the 2019 Structured Preapproval and Postapproval Comparative study design framework to generate valid and transparent real‐world Evidence (SPACE) framework elucidated a process for designing valid and transparent real‐world studies. As an extension to SPACE, here, we provide a structured framework for conducting feasibility assessments—a step‐by‐step guide to identify decision grade, fit‐for‐purpose data, which complements the United States Food and Drug Administration (FDA)’s framework for a RWE program. The process was informed by our collective experience conducting systematic feasibility assessments of existing data sources for pharmacoepidemiology studies to support regulatory decisions. Used with the SPACE framework, the Structured Process to Identify Fit‐For‐Purpose Data (SPIFD) provides a systematic process for conducting feasibility assessments to determine if a data source is fit for decision making, helping ensure justification and transparency throughout study development, from articulation of a specific and meaningful research question to identification of fit‐for‐purpose data and study design.

Real-world evidence: a practical toolbox for collecting health state utilities

Health state utilities (HSU) data collected in real-world evidence studies are at risk of bias. Although numerous guidance documents are available, practical advice to avoid bias in HSU studies is limited. Thus, the objective of this article was to develop a concise toolbox intended for investigators seeking to collect HSU in a real-world setting. The proposed toolbox builds on existing guidance and provides practical steps to help investigators perform good quality research. The toolbox aims at increasing the credibility of HSU data for future reimbursement decision making.

Differences in Antipsychotic Treatment Discontinuation Among Veterans With Schizophrenia in the U.S. Department of Veterans Affairs

OBJECTIVE

Effectiveness of antipsychotic drugs is inferred from relatively small randomized clinical trials conducted with carefully selected and monitored participants. This evidence is not necessarily generalizable to individuals treated in daily clinical practice. The authors compared the clinical effectiveness between all oral and long-acting injectable (LAI) antipsychotic medications used in the treatment of schizophrenia in the U.S. Department of Veterans Affairs (VA) health care system.

METHODS

This was an observational study utilizing VA pharmacy data from 37,368 outpatient veterans with schizophrenia. Outcome measures were all-cause antipsychotic discontinuation and psychiatric hospitalizations. Oral olanzapine was used as the reference group.

RESULTS

In multivariable analysis, clozapine (hazard ratio=0.43), aripiprazole long-acting injectable (LAI) (hazard ratio=0.71), paliperidone LAI (hazard ratio=0.76), antipsychotic polypharmacy (hazard ratio=0.77), and risperidone LAI (hazard ratio=0.91) were associated with reduced hazard of discontinuation compared with oral olanzapine. Oral first-generation antipsychotics (hazard ratio=1.16), oral risperidone (hazard ratio=1.15), oral aripiprazole (hazard ratio=1.14), oral ziprasidone (hazard ratio=1.13), and oral quetiapine (hazard ratio=1.11) were significantly associated with an increased risk of discontinuation compared with oral olanzapine. No treatment showed reduced risk of psychiatric hospitalization compared with oral olanzapine; quetiapine was associated with a 36% worse outcome in terms of hospitalizations compared with olanzapine.

CONCLUSIONS

In a national sample of veterans with schizophrenia, those treated with clozapine, two of the LAI second-generation antipsychotics, and antipsychotic polypharmacy continued the same antipsychotic therapy for a longer period of time compared with the reference drug. This may reflect greater overall acceptability of these medications in clinical practice.

The anemia-independent impact of myelodysplastic syndromes on health-related quality of life

Myelodysplastic syndromes (MDS) are in the majority of cases characterized by anemia. Both anemia and MDS per se may directly contribute to impairments in health-related quality of life (HRQoL). In this study, we aimed to investigate the anemia-independent impact of MDS on HRQoL. We evaluated participants (≥ 50 years) from the large population-based Lifelines cohort (N = 44,694, mean age 59.0 ± 7.4 years, 43.6% male) and the European MDS Registry (EUMDS) (N = 1538, mean age 73.4 ± 9.0 years, 63.0% male), which comprises a cohort of lower-risk MDS patients. To enable comparison concerning HRQoL, SF-36 scores measured in Lifelines were converted to EQ-5D-3L index (range 0–1) and dimension scores. Lower-risk MDS patients had significantly lower HRQoL than those from the Lifelines cohort, as illustrated in both the index score and in the five different dimensions. Multivariable linear regression analysis demonstrated that MDS had an adjusted total impact on the EQ-5D index score (B =  − 0.12, p < 0.001) and an anemia-independent “direct” impact (B =  − 0.10, p < 0.001). Multivariable logistic regression analysis revealed an anemia-independent impact of MDS in the dimension mobility, self-care, usual activities, and anxiety/depression (all except pain/discomfort). This study demonstrates that the major part of the negative impact of lower-risk MDS on HRQoL is not mediated via anemia. Thus, the therapeutic focus should include treatment strategies directed at underlying pathogenic mechanisms to improve HRQoL, rather than aiming predominantly at increasing hemoglobin levels.

Secondary Obliteration Surgery for Troublesome Mastoid Cavities: A Prospective Series of 228 Pediatric and Adult Cases

OBJECTIVE

To present the results of a 15-year longitudinal study in a pediatric and adult population with secondary obliteration of troublesome mastoid cavities.

STUDY DESIGN

Prospective longitudinal study.

PATIENTS

Children (≤16 yrs) and adults who had obliteration surgery (228 ears) between 2000 and 2017.

INTERVENTIONS

Therapeutic.

SETTING

Tertiary referral center.

MAIN OUTCOME MEASURES

1) Incidence of recurrent or residual cholesteatoma; 2) creation of a dry ear at 5 years postsurgery; 3) postoperative complications; 4) waterproof status of ear; 5) number of subsequent ear surgery required.

RESULTS

Thirty-one pediatric and 197 adult cases contributed to the study. At 5 years 11 of 98 (10.8%) had been lost to follow-up. Using Kaplan-Meier survival analysis, the residual rate after 5 years was 4.2% (95% CI: 0.3%-8.2%) and the recurrence rate was 1% (95% CI: 0.0%-3.0%). Of the 223 ears available for assessment, 209 (93.7%) achieved a dry state, 6 (2.7%) had intermittent discharge, 6 (2.7%) had flap necrosis requiring flap trimming, and 1 (0.4%) developed meatal stenosis. Using a cross-sectional analysis at 12 months of follow-up, the otorrhea risk was 19.2% and the risk of definitive waterproofing was 10.9%. There was a reoperation risk of 10% within 5 years which included second-stage ossiculoplasty.

CONCLUSIONS

Secondary mastoid obliteration is a safe and useful technique in treating the troublesome mastoid cavity in both children and adults. It is associated with a low cholesteatoma recidivism rate and high rate of a trouble-free ear in the long term.

Electronic Medical Record Risk Modeling of Cardiovascular Outcomes Among Patients with Type 2 Diabetes

INTRODUCTION

Increased utilization of electronic health records (EHR) has enriched databases for creating risk models. We used machine learning techniques to develop an EHR-based risk model locally fitted to patients with type 2 diabetes mellitus (T2DM) for predicting cardiovascular disease.

METHODS

This retrospective observational study was conducted within Ochsner Health, Louisiana, USA, between 2013-2017. Data analysis included 6245 patients who had two outpatient diagnoses of T2DM recorded on separate days or a diagnosis recorded during an inpatient encounter. Baseline clinical data were limited to 180 days before the index diagnosis. Cardiovascular outcomes were coronary heart disease (CHD), heart failure and stroke. Machine learning approaches were used to select predictor variables into Cox proportional hazards models for each outcome. Locally fit equations were compared to "generalized" risk equations (RECODe, AS-CVD, QRISK3) using model discrimination and calibration.

RESULTS

Among factors identified in the Ochsner (n = 11), RECODe (n = 14), AS-CVD (n = 15) and QRISK3 (n = 23), only age was common to all four risk equations. The Ochsner model had high internal discrimination for CHD (C-statistics 0.85) and better discrimination than RECODe (C-statistics 0.45), the QRISK3 (C-statistics 0.72) and AS-CVD (C-statistics 0.54).

CONCLUSIONS

The Ochsner model overestimated 5-year CHD risk, but had relatively higher calibration than the other models in CHD. Risk equations fitted for local populations improved cardiovascular risk stratification for patients with T2DM. Application of machine learning simplified the models compared to "generalized" risk equations.

New-user and prevalent-user designs and the definition of study time origin in pharmacoepidemiology: A review of reporting practices

BACKGROUND

Guidance reports for observational comparative effectiveness and drug safety research recommend implementing a new-user design whenever possible, since it reduces the risk of selection bias in exposure effect estimation compared to a prevalent-user design. The uptake of this guidance has not been studied extensively.

METHODS

We reviewed 89 observational effectiveness and safety cohort studies published in six pharmacoepidemiological journals in 2018 and 2019. We developed an extraction tool to assess how frequently new-user and prevalent-user designs were reported to be implemented. For studies that implemented a new-user design in both treatment arms, we extracted information about the extent to which the moment of meeting eligibility criteria, treatment initiation, and start of follow-up were reported to be aligned.

RESULTS

Of the 89 studies included, 40% reported implementing a new-user design for both the study exposure arm and the comparator arm, while 13% reported implementing a prevalent-user design in both arms. The moment of meeting eligibility criteria, treatment initiation, and start of follow-up were reported to be aligned in both treatment arms in 53% of studies that reported implementing a new-user design. We provided examples of studies that minimized the risk of introducing bias due to unclear definition of time origin in unexposed participants, immortal time, or a time lag.

CONCLUSIONS

Almost half of the included studies reported implementing a new-user design. Implications of misalignment of study design origin were difficult to assess because it would require explicit reporting of the target estimand in original studies. We recommend that the choice for a particular study time origin is explicitly motivated to enable assessment of validity of the study.

Development of An Individualized Risk Prediction Model for COVID-19 Using Electronic Health Record Data

Developing an accurate and interpretable model to predict an individual's risk for Coronavirus Disease 2019 (COVID-19) is a critical step to efficiently triage testing and other scarce preventative resources. To aid in this effort, we have developed an interpretable risk calculator that utilized de-identified electronic health records (EHR) from the University of Alabama at Birmingham Informatics for Integrating Biology and the Bedside (UAB-i2b2) COVID-19 repository under the U-BRITE framework. The generated risk scores are analogous to commonly used credit scores where higher scores indicate higher risks for COVID-19 infection. By design, these risk scores can easily be calculated in spreadsheets or even with pen and paper. To predict risk, we implemented a Credit Scorecard modeling approach on longitudinal EHR data from 7,262 patients enrolled in the UAB Health System who were evaluated and/or tested for COVID-19 between January and June 2020. In this cohort, 912 patients were positive for COVID-19. Our workflow considered the timing of symptoms and medical conditions and tested the effects by applying different variable selection techniques such as LASSO and Elastic-Net. Within the two weeks before a COVID-19 diagnosis, the most predictive features were respiratory symptoms such as cough, abnormalities of breathing, pain in the throat and chest as well as other chronic conditions including nicotine dependence and major depressive disorder. When extending the timeframe to include all medical conditions across all time, our models also uncovered several chronic conditions impacting the respiratory, cardiovascular, central nervous and urinary organ systems. The whole pipeline of data processing, risk modeling and web-based risk calculator can be applied to any EHR data following the OMOP common data format. The results can be employed to generate questionnaires to estimate COVID-19 risk for screening in building entries or to optimize hospital resources.

Organized structure of real-world evidence best practices: moving from fragmented recommendations to comprehensive guidance

Decision-makers have become increasingly interested in incorporating real-world evidence (RWE) into their decision-making process. Due to concerns regarding the reliability and quality of RWE, stakeholders have issued numerous recommendation documents to assist in setting RWE standards. The fragmented nature of these documents poses a challenge to researchers and decision-makers looking for guidance on what is 'high-quality' RWE and how it can be used in decision-making. We offer researchers and decision-makers a structure to organize the landscape of RWE recommendations and identify consensus and gaps in the current recommendations. To provide researchers with a much needed pathway for generating RWE, we discuss how decision-makers can move from fragmented recommendations to comprehensive guidance.

Impact of anticoagulant exposure misclassification on the bleeding risk of direct oral anticoagulants

Aims

Drug exposure status based on routinely collected data might be misclassified when the database contains only prescriptions from 1 type of prescriber (e.g. general practitioner and not specialist). This study aims to quantify the impact of such exposure misclassification on the risk of major bleeding and stroke/transient ischaemic attack (TIA)associated with direct oral anticoagulants (DOACs) vs. vitamin K antagonists (VKAs).

Methods

Incident anticoagulant users (>12 mo free of anticoagulation use) in the Dutch PHARMO Database Network between 2008 and 2017 were included. Drug exposure was assessed using pharmacy dispensing information. The risks of hospital admission of major bleeding for DOAC vs. VKA users was assessed with Cox regression analysis, where exposure was based on all dispensings, on general practitioner (GP)‐prescribed dispensings only or on specialist‐prescribed dispensings only. Hazard ratios (HRs) were estimated also for hospitalization for gastrointestinal bleeding, intracranial bleeding and stroke/TIA.

Results

We included 99 182 VKA‐initiators and 21 795 DOAC‐initiators. Use of DOAC was associated with a lower risk of major bleeding compared to VKA use; HR 0.79 (95% confidence interval 0.70–0.90), 0.78 (0.68–0.91) and 0.62 (0.50–0.76), for exposure based on complete dispensing information, only GP‐ and only specialist‐prescribed dispensings, respectively. Similar results were found for the other bleeding outcomes. For stroke/TIA the HRs were 0.96 (0.84–1.09), 1.00 (0.84–1.18) and 0.72 (0.58–0.90), respectively.

Conclusion

Including only GP‐prescribed anticoagulant dispensings in this case did not materially impact the effect estimates compared to including all anticoagulant dispensings. Including only specialist‐prescribed dispensings, however, strengthened the effect estimates.

Mexican physical education teacher’s self-reported and self-perception of teaching style

ABSTRACT To understand the self-reported and perception of teaching style used among Mexican physical education teachers. Participants included 104 (n=41 females and 63 males) elementary physical education teachers from northern Mexico. Results showed that a traditional teaching style tends to be mainly used either by male and female’s physical education teachers. Showing differences between genders and teaching styles. Female teachers tend to combine a traditional teaching style with one that connects with students engaging in physical activity, developing positive values and attitudes. Cultural context plays an important aspect when developing a teaching self-perception.

Identifying naloxone administrations in electronic health record data using a text-mining tool

Background: Effective and efficient methods are needed to identify naloxone administrations within electronic health record (EHR) data to conduct overdose surveillance and research. The objective of this study was to develop and validate a text-mining tool to identify naloxone administrations in EHR data. Methods: Clinical notes stored in databases between January 2017 and March 2018 were used to iteratively develop a text-mining tool to identify naloxone administrations. The first iteration of the tool used broad search terms. Then, after reviewing clinical notes of overdose encounters, we developed a list of phrases that described naloxone administrations to inform iteration two. While validating iteration two, additional phrases were found, which were then added to inform the final iteration. The comparator was an administrative code query extracted from the EHR. Medical record review was used to identify true positives. The primary outcome was the positive predictive values (PPV) of the second iteration, final iteration, and administrative code query. Results: Iteration two, the final iteration, and the administrative code had PPVs of 84.3% (95% confidence interval [CI] 78.6-89.0%), 83.8% (95% CI 78.6-88.2%), and 57.1% (95% CI 47.1-66.8%), respectively. Both iterations of the tool had a significantly higher PPV than the administrative code (p < 0.001). Conclusions: A text-mining tool improved the identification of naloxone administrations in EHR data from less than 60% with the administrative code to greater than 80% with both versions of the tool. Text-mining tools can inform the use of more sophisticated informatics methods, which often require significant time, resource, and expertise investment.

Comparison of pleiotropic effects of statins vs fibrates on laboratory parameters in patients with dyslipidemia: A retrospective cohort study

Differences in the mechanism of action and potential pleiotropic effects between statins and fibrates would potentially drive a different effect on various laboratory parameters, but this remains controversial because of a paucity of reports comparing them. Therefore, the aim of this study was to compare the effects of statins and fibrates on laboratory parameters in Japanese patients in routine clinical practice.This retrospective cohort study included patients with dyslipidemia who had been newly treated with statin or fibrate monotherapy between January 2005 and December 2017. Patients were randomly matched into two sets of pairs by sex, age, and baseline triglyceride (TG) or low-density lipoprotein (LDL) cholesterol level. The 830 patients in TG-matched pairs (415 fibrate users and 415 matched statin users) and 1172 patients in LDL cholesterol-matched pairs (586 fibrate users and 586 matched statin users) were included in this study. Generalized estimating equations were used to estimate the effects of the drugs on serum creatinine level, estimated glomerular filtration rate (eGFR), urea nitrogen, hemoglobin A1c, aspartate aminotransferase, and alanine aminotransferase (ALT), in addition to LDL cholesterol and TG levels, and red blood cell (RBC) and platelet (PLT) counts, up to 12 months after the start of study drug administration.In TG-matched pairs, the increases in creatinine and urea nitrogen levels (P = .010 and P < .001, respectively) and the decreases in eGFR, ALT level and RBC count (P < .001, P = .003, and P = .014, respectively) were greater in fibrate users than in statin users. The decrease in PLT count was greater in statin users than in fibrate users (P < .001). The mean changes in aspartate aminotransferase and hemoglobin A1c levels were not significantly different between statin users and fibrate users. In LDL cholesterol-matched pairs, the differences in changes of all laboratory parameter levels between statin users and fibrate users were similar to those in TG-matched pairs.We demonstrate here that fibrates have a greater effect of increasing creatinine and urea nitrogen levels and of reducing eGFR, ALT level, and RBC count than statins, and that the lowering effect on PLT count is greater with statins than with fibrates.

Is Real-World Evidence Used in P&T Monographs and Therapeutic Class Reviews?

BACKGROUND

Payers are faced with making coverage and reimbursement decisions based on the best available evidence. Often these decisions apply to patient populations, provider networks, and care settings not typically studied in clinical trials. Treatment effectiveness evidence is increasingly available from electronic health records, registries, and administrative claims. However, little is known about when and what types of real-world evidence (RWE) studies inform pharmacy and therapeutic (P&T) committee decisions.

OBJECTIVE

To evaluate evidence sources cited in P&T committee monographs and therapeutic class reviews and assess the design features and quality of cited RWE studies.

METHODS

A convenience sample of representatives from pharmacy benefit management, health system, and health plan organizations provided recent P&T monographs and therapeutic class reviews (or references from such documents). Two investigators examined and grouped references into major categories (published studies, unpublished studies, and other/unknown) and multiple subcategories (e.g., product label, clinical trials, RWE, systematic reviews). Cited comparative RWE was reviewed to assess design features (e.g., population, data source, comparators) and quality using the Good ReseArch for Comparative Effectiveness (GRACE) Checklist.

RESULTS

Investigators evaluated 565 references cited in 27 monographs/therapeutic class reviews from 6 managed care organizations. Therapeutic class reviews mostly cited published clinical trials (35.3%, 155/439), while single-product monographs relied most on manufacturer-supplied information (42.1%, 53/126). Published RWE comprised 4.8% (21/439) of therapeutic class review references, and none (0/126) of the monograph references. Of the 21 RWE studies, 12 were comparative and assessed patient care settings and outcomes typically not included in clinical trials (community ambulatory settings [10], long-term safety [8]). RWE studies most frequently were based on registry data (6), conducted in the United States (6), and funded by the pharmaceutical industry (5). GRACE Checklist ratings suggested the data and methods of these comparative RWE studies were of high quality.

CONCLUSIONS

RWE was infrequently cited in P&T materials, even among therapeutic class reviews where RWE is more readily available. Although few P&T materials cited RWE, the comparative RWE studies were generally high quality. More research is needed to understand when and what types of real-world studies can more routinely inform coverage and reimbursement decisions.

DISCLOSURES

This project was funded by the National Pharmaceutical Council. Hurwitz, Brown, Peters, and Malone have nothing to disclose. Graff is employed by the National Pharmaceutical Council Part of this study was presented as a poster presentation at the AMCP Managed Care & Specialty Pharmacy 2016 Annual Meeting; April 19-22, 2016; San Francisco, CA. Study concept and design were primarily contributed by Malone and Graff, along with Hurwitz and Brown. All authors participated in data collection, and data interpretation was performed by Malone, Hurwitz, and Graff, with assistance from Brown and Peters. The manuscript was written primarily by Hurwitz and Malone, along with Graff, Brown, and Peters, and revised by Malone, Brown, Peters, Hurwitz, and Graff.

Assessment of Studies Evaluating Incremental Costs, Effectiveness, or Cost-Effectiveness of Systemic Therapies in Breast Cancer Based on Claims Data: A Systematic Review

OBJECTIVES

Large secondary databases, such as those containing insurance claims data, are increasingly being used to compare the effects and costs of treatments in routine clinical practice. Despite their appeal, however, caution must be exercised when using these data. In this study, we aimed to identify and assess the methodological quality of studies that used claims data to compare the effectiveness, costs, or cost-effectiveness of systemic therapies for breast cancer.

METHODS

We searched Embase, the Cochrane Library, Medline, Web of Science, and Google Scholar for English-language publications and assessed the methodological quality using the Good Research for Comparative Effectiveness principles. This study was registered with the International Prospective Register of Systematic Reviews (PROSPERO) under number CRD42018103992.

RESULTS

We identified 1251 articles, of which 106 met the inclusion criteria. Most studies were conducted in the United States (74%) and Taiwan (9%) and were based on claims data sets (35%) or claims data linked to cancer registries (58%). Furthermore, most included large samples (mean 17 130 patients) and elderly patients, and they covered various outcomes (eg, survival, adverse events, resource use, and costs). Key methodological shortcomings were the lack of information on relevant confounders, the risk of immortal time bias, and the lack of information on the validity of outcomes. Only a few studies performed sensitivity analyses.

CONCLUSIONS

Many comparative studies of cost, effectiveness, and cost-effectiveness have been published in recent decades based on claims data, and the number of publications has increased over time. Despite the availability of guidelines to improve quality, methodological issues persist and are often inappropriately addressed or reported.

The Impact of Intestinal Complications on Health Care Costs Among Patients With Inflammatory Bowel Disease Treated With Anti-Tumor Necrosis Factor Therapies

BACKGROUND

Although there is evidence that anti-tumor necrosis factor (TNF) utilization earlier in the inflammatory bowel disease (IBD) course and before the onset of disease-related complications leads to improved patient outcomes, the health care costs and utilization impact have not been well defined. This study assessed differences in health care utilization and costs among patients with IBD treated with anti-TNFs.

METHODS

Patients with a diagnosis of ulcerative colitis (UC) or Crohn disease (CD) between January 1, 2001, and December 31, 2014, were identified from a claims database. Patients were required to have ≥1 claim for a 5-aminosalicylic acid, corticosteroid, or immunomodulator after the IBD diagnosis and ≥1 anti-TNF drug claim after the first IBD treatment. Complication and noncomplication cohorts were identified based on disease-related complications and IBD-related hospitalizations or emergency department visits for 6 months before anti-TNF initiation. Generalized linear models were used to compare health care costs and utilization for the 12 months after anti-TNF initiation (follow-up).

RESULTS

The study included 6329 patients with CD and 4451 patients with UC. In patients with CD with complications, >33.7% had intestinal strictures and 6% had enteroenteric fistula before anti-TNF treatment. Patients with CD with complications incurred significantly higher IBD-related and all-cause health care costs during follow-up, and patients with UC experienced the same trends.

CONCLUSIONS

These results suggest that anti-TNF treatment after, rather than before, a patient develops complications leads to a higher economic burden. However, these findings could also result from patients with more severe disease having early complications that are more difficult to treat.

Validation of venous thromboembolism diagnoses in patients receiving rivaroxaban or warfarin in The Health Improvement Network

Purpose

To describe the effect that validation of venous thromboembolism (VTE) coded entries in the health improvement network (THIN) has on incidence rates of VTE among a cohort of rivaroxaban/warfarin users.

Methods

Among 36 701 individuals with a first prescription for rivaroxaban/warfarin between 2012 and 2015, we performed a two‐step VTE case identification process followed by a two‐step case validation process involving manual review of patient records. A valid case required a coded entry for VTE at some point after their first rivaroxaban/warfarin prescription with evidence of referral/hospitalization either as a coded entry or entered as free text. Positive predictive values (PPVs) with 95% confidence intervals (CIs) were calculated using validated cases as the gold standard. Incidence rates were calculated per 1000 person‐years with 95% CIs.

Results

We identified 2166 patients with a coded entry of VTE after their initial rivaroxaban/warfarin prescription; incidence rate of 45.31 per 1000 person‐years (95% CI: 43.49‐47.22). After manual review of patient records including the free text, there were 712 incident VTE cases; incidence rate of 14.90 per 1000 person‐years (95% CI: 13.85‐16.02). The PPV for coded entries of VTE alone was 32.9%, and the PPV for coded entries of VTE with a coded entry of referral/hospitalization was 39.8%; this increased to 69.6% after manual review of coded clinical entries in patient records.

Conclusions

Among rivaroxaban/warfarin users in THIN, valid VTE case identification requires manual review of patient records including the free text to prevent outcome misclassification and substantial overestimation of VTE incidence rates.

Health-related quality of life, work productivity and costs related to patients with inflammatory bowel disease in Austria

OBJECTIVE

Inflammatory-Bowel-Disease (IBD) is a lifelong illness with significant impact on health-related quality of life (HRQoL). The disease-burden causes work productivity impairment, such as sick-leave and restriction of leisure time activities. From a societal perspective, productivity loss often contributes significantly to the total costs. The aim of the study is to analyze the impact of disease-burden on work productivity, daily activities, and HRQoL.

METHODS

We performed a survey among Austrian patients with IBD to assess general questions about demography and disease pattern, impaired work productivity and reduced daily activities using the Work Productivity and Activity Impairment questionnaire and HRQoL utilizing the Short Inflammatory Bowel Disease Questionnaire (SIBDQ). Associations were assessed employing linear regressions. The questionnaire was circulated by the Austrian Crohn's Disease/Ulcerative Colitis Association.

RESULTS

In total, 510 patients with median age of 40 years completed the questionnaire. Patients had a median disease-duration of 9 years, 46% were treated with biologics and 64% were employed. The median SIBDQ score was 45 (16-69). Patients in self-assessed remission (n = 319) and relapse (n = 96) differed by 18 points on the SIBDQ score (p value < 0.0001). Mean work productivity reduction amounted to 7 h 17 min (95% CI 6:05-8:29) during the last 7 days. Work absenteeism accounted for 4 h 28 min (95% CI 3:25-5:31) and reduced productivity at the workplace (presenteeism) for 2 h 49 min (95% CI 2:31-3:07). During a relapse, 12 h 9 min were lost (p value < 0.0001). Productivity reduction correlated with HRQoL (r = 0.3964). In total, 143 €per patient per week (95% CI 119-167) were lost (7,411 €per year). On average, patients lost 4 h (95% CI 3:44-4:16) of unpaid daily activities; 3 h 17 min the employed subgroup and 5 h 19 min the not employed.

CONCLUSION

IBD leads to a substantial burden of disease, which is associated with indirect costs for society related to work ability reduction. HRQoL significantly influences loss of work productivity. Improving HRQoL should be a key factor to reduce disease burden and societal costs.

Improving transparency to build trust in real-world secondary data studies for hypothesis testing-Why, what, and how: recommendations and a road map from the real-world evidence transparency initiative

Real-world data (RWD) and the derivations of these data into real-world evidence (RWE) are rapidly expanding from informing healthcare decisions at the patient and health system level to influencing major health policy decisions, including regulatory approvals and coverage. Recent examples include the approval of palbociclib in combination with endocrine therapy for male breast cancer and the inclusion of RWE in the label of paliperidone palmitate for schizophrenia. This interest has created an urgency to develop processes that promote trust in the evidence-generation process. Key stakeholders and decision-makers include patients and their healthcare providers; learning health systems; health technology assessment bodies and payers; pharmacoepidemiologists and other clinical reseachers, and policy makers interested in bioethical and regulatory issues. A key to optimal uptake of RWE is transparency of the research process to enable decision-makers to evaluate the quality of the methods used and the applicability of the evidence that results from the RWE studies. Registration of RWE studies-particularly for hypothesis evaluating treatment effectiveness (HETE) studies-has been proposed to improve transparency, trust, and research replicability. Although registration would not guarantee better RWE studies would be conducted, it would encourage the prospective disclosure of study plans, timing, and rationale for modifications. A joint task force of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) and the International Society for Pharmacoepidemiology (ISPE) recommended that investigators preregister their RWE studies and post their study protocols in a publicly available forum before starting studies to reduce publication bias and improve the transparency of research methods. Recognizing that published recommendations alone are insufficient, especially without accessible registration options and with no incentives, a group of experts gathered on February 25 and 26, 2019, in National Harbor, Maryland, to explore the structural and practical challenges to the successful implementation of the recommendations of the ISPOR/ISPE task force for preregistration. This positioning article describes a plan for making registration of HETE RWE studies routine. The plan includes specifying the rationale for registering HETE RWE studies, the studies that should be registered, where and when these studies should be registered, how and when analytic deviations from protocols should be reported, how and when to publish results, and incentives to encourage registration. Table 1 summarizes the rationale, goals, and potential solutions that increase transparency, in addition to unique concerns about secondary data studies. Definitions of terms used throughout this report are provided in Table 2.

Real-world evidence: the devil is in the detail

Much has been written about real-world evidence (RWE), a concept that offers an understanding of the effects of healthcare interventions using routine clinical data. The reflection of diverse real-world practices is a double-edged sword that makes RWE attractive but also opens doors to several biases that need to be minimised both in the design and analytical phases of non-experimental studies. Additionally, it is critical to ensure that researchers who conduct these studies possess adequate methodological expertise and ability to accurately implement these methods. Critical design elements to be considered should include a clearly defined research question using a causal inference framework, choice of a fit-for-purpose data source, inclusion of new users of a treatment with comparators that are as similar as possible to that group, accurately classifying person-time and deciding censoring approaches. Having taken measures to minimise bias 'by design', the next step is to implement appropriate analytical techniques (for example propensity scores) to minimise the remnant potential biases. A clear protocol should be provided at the beginning of the study and a report of the results after, including caveats to consider. We also point the readers to readings on some novel analytical methods as well as newer areas of application of RWE. While there is no one-size-fits-all solution to evaluating RWE studies, we have focused our discussion on key methods and issues commonly encountered in comparative observational cohort studies with the hope that readers are better equipped to evaluate non-experimental studies that they encounter in the future. Graphical abstract.

Improving Transparency to Build Trust in Real-World Secondary Data Studies for Hypothesis Testing-Why, What, and How: Recommendations and a Road Map from the Real-World Evidence Transparency Initiative

Real-world data (RWD) and the derivations of these data into real-world evidence (RWE) are rapidly expanding from informing healthcare decisions at the patient and health system level to influencing major health policy decisions, including regulatory approvals and coverage. Recent examples include the approval of palbociclib in combination with endocrine therapy for male breast cancer and the inclusion of RWE in the label of paliperidone palmitate for schizophrenia. This interest has created an urgency to develop processes that promote trust in the evidence-generation process. Key stakeholders and decision-makers include patients and their healthcare providers; learning health systems; health technology assessment bodies and payers; pharmacoepidemiologists and other clinical reseachers, and policy makers interested in bioethical and regulatory issues. A key to optimal uptake of RWE is transparency of the research process to enable decision-makers to evaluate the quality of the methods used and the applicability of the evidence that results from the RWE studies. Registration of RWE studies-particularly for hypothesis evaluating treatment effectiveness (HETE) studies-has been proposed to improve transparency, trust, and research replicability. Although registration would not guarantee better RWE studies would be conducted, it would encourage the prospective disclosure of study plans, timing, and rationale for modifications. A joint task force of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) and the International Society for Pharmacoepidemiology (ISPE) recommended that investigators preregister their RWE studies and post their study protocols in a publicly available forum before starting studies to reduce publication bias and improve the transparency of research methods. Recognizing that published recommendations alone are insufficient, especially without accessible registration options and with no incentives, a group of experts gathered on February 25 and 26, 2019, in National Harbor, Maryland, to explore the structural and practical challenges to the successful implementation of the recommendations of the ISPOR/ISPE task force for preregistration. This positioning article describes a plan for making registration of HETE RWE studies routine. The plan includes specifying the rationale for registering HETE RWE studies, the studies that should be registered, where and when these studies should be registered, how and when analytic deviations from protocols should be reported, how and when to publish results, and incentives to encourage registration. Table 1 summarizes the rationale, goals, and potential solutions that increase transparency, in addition to unique concerns about secondary data studies. Definitions of terms used throughout this report are provided in Table 2.

Does the Encounter Type Matter When Defining Diabetes Complications in Electronic Health Records?

BACKGROUND

Electronic health records (EHRs) and claims records are widely used in defining type 2 diabetes mellitus (T2DM) complications across different types of health care encounters.

OBJECTIVE

This study investigates whether using different EHR encounter types to define diabetes complications may lead to different results when examining associations between diabetes complications and their risk factors in patients with T2DM.

RESEARCH DESIGN

The study cohort of 64,855 adult patients with T2DM was created from EHR data from the Research Action for Health Network (REACHnet), using the Surveillance Prevention, and Management of Diabetes Mellitus (SUPREME-DM) definitions. Incidence of coronary heart disease (CHD) and stroke events were identified using International Classification of Diseases (ICD)-9/10 codes and grouped by encounter types: (1) inpatient (IP) or emergency department (ED) type, or (2) any health care encounter type. Cox proportional hazards regression was used to estimate associations between diabetes complications (ie, CHD and stroke) and risk factors (ie, low-density lipoprotein cholesterol and hemoglobin A1c).

RESULTS

The incidence rates of CHD and stroke in all health care settings were more than twice the incidence rates of CHD and stroke in IP/ED settings. The age-adjusted and multivariable-adjusted hazard ratios for incident CHD and stroke across different levels of low-density lipoprotein cholesterol and hemoglobin A1c were similar between IP/ED and all settings.

CONCLUSION

While there are large variations in incidence rates of CHD and stroke as absolute risks, the associations between both CHD and stroke and their respective risk factors measured by hazard ratios as relative risks are similar, regardless of alternative definitions.

Reliability of Conclusions from Early Analyses of Real-World Data for Newly Approved Drugs in Advanced Gastric Cancer in the United States

Background

As real-world data resources expand and improve, there will increasingly be opportunities to study the effectiveness of interventions. There is a need to ensure that study designs explore potential sources of bias and either acknowledge or mitigate them, in order to improve the accuracy of findings. The objective of this study was to understand newly approved drug utilization patterns in real-world clinical settings over time.

Methods

This retrospective study included three sources of real-world data (claims, electronic health records, and recoded data from a quality care program) collected from patients diagnosed with gastric cancer who initiated therapy with either trastuzumab or ramucirumab. Linear regression was used to investigate trends in the use of these drugs for the care of patients with gastric cancer over time from Food and Drug Administration (FDA) approval.

Results

Eligible patients (n=1700) had consistent demographic and clinical characteristics over time. After regulatory approval, trastuzumab was used in later lines of therapy and then shifted to earlier lines (p=0.002), while ramucirumab utilization remained consistent over time after FDA approval (p=0.49). Ramucirumab augmentation, defined as the addition of the drug after initiation of a line of therapy, decreased over time (p=0.03), and trastuzumab augmentation remained consistent over time (p=0.58).

Conclusion

Since treatment effectiveness may change across lines of treatment, bias may arise if there are changes in the use of the drug (such as line migration) during the time period of analysis using real-world data.

Association Between Initiation of Pulmonary Rehabilitation After Hospitalization for COPD and 1-Year Survival Among Medicare Beneficiaries

IMPORTANCE

Meta-analyses have suggested that initiating pulmonary rehabilitation after an exacerbation of chronic obstructive pulmonary disease (COPD) was associated with improved survival, although the number of patients studied was small and heterogeneity was high. Current guidelines recommend that patients enroll in pulmonary rehabilitation after hospital discharge.

OBJECTIVE

To determine the association between the initiation of pulmonary rehabilitation within 90 days of hospital discharge and 1-year survival.

DESIGN, SETTING, AND PATIENTS

This retrospective, inception cohort study used claims data from fee-for-service Medicare beneficiaries hospitalized for COPD in 2014, at 4446 acute care hospitals in the US. The final date of follow-up was December 31, 2015.

EXPOSURES

Initiation of pulmonary rehabilitation within 90 days of hospital discharge.

MAIN OUTCOMES AND MEASURES

The primary outcome was all-cause mortality at 1 year. Time from discharge to death was modeled using Cox regression with time-varying exposure to pulmonary rehabilitation, adjusting for mortality and for unbalanced characteristics and propensity to initiate pulmonary rehabilitation. Additional analyses evaluated the association between timing of pulmonary rehabilitation and mortality and between number of sessions completed and mortality.

RESULTS

Of 197 376 patients (mean age, 76.9 years; 115 690 [58.6%] women), 2721 (1.5%) initiated pulmonary rehabilitation within 90 days of discharge. A total of 38 302 (19.4%) died within 1 year of discharge, including 7.3% of patients who initiated pulmonary rehabilitation within 90 days and 19.6% of patients who initiated pulmonary rehabilitation after 90 days or not at all. Initiation within 90 days was significantly associated with lower risk of death over 1 year (absolute risk difference [ARD], -6.7% [95% CI, -7.9% to -5.6%]; hazard ratio [HR], 0.63 [95% CI, 0.57 to 0.69]; P < .001). Initiation of pulmonary rehabilitation was significantly associated with lower mortality across start dates ranging from 30 days or less (ARD, -4.6% [95% CI, -5.9% to -3.2%]; HR, 0.74 [95% CI, 0.67 to 0.82]; P < .001) to 61 to 90 days after discharge (ARD, -11.1% [95% CI, -13.2% to -8.4%]; HR, 0.40 [95% CI, 0.30 to 0.54]; P < .001). Every 3 additional sessions was significantly associated with lower risk of death (HR, 0.91 [95% CI, 0.85 to 0.98]; P = .01).

CONCLUSIONS AND RELEVANCE

Among fee-for-service Medicare beneficiaries hospitalized for COPD, initiation of pulmonary rehabilitation within 3 months of discharge was significantly associated with lower risk of mortality at 1 year. These findings support current guideline recommendations for pulmonary rehabilitation after hospitalization for COPD, although the potential for residual confounding exists and further research is needed.