Use of health insurance claim patterns to identify patients using nonsteroidal anti-inflammatory drugs for rheumatoid arthritis

Suboptimal management of rheumatoid arthritis in France: a real-world study based on data from the French National Health Data System

OBJECTIVES

The emergence of targeted therapy is changing rheumatoid arthritis (RA) management, but real-world data remain limited. This study aimed to describe real-world RA treatment patterns using data from a French national claims database.

METHODS

This longitudinal study used the French Permanent Representative Sample (Echantillon Généraliste des Bénéficiaires) claims database. Patients with RA were identified between 2013 and 2017, with treatment patterns, persistence and adherence described.

RESULTS

The study population included 2553 patients with RA. Disease-modifying antirheumatic drugs (DMARDs) were prescribed for 1512 (59.2%) patients, of whom 721 (47.6%) did not require discontinuation or treatment switch. There were 377 (24.9%) treatment discontinuations and 114 patients (7.5%) switched to a targeted DMARD (biological and synthetic (Janus kinase inhibitor) DMARDs). Among the 2315 patients with RA in 2017, almost half (n=1102, 47.6%) were not treated with a DMARD. Most (85.7%) received symptomatic treatment (analgesics (81.0%), steroids (49.2%), non-steroidal anti-inflammatory drugs (39.5%)). Of the 1142 treatment initiations identified, 713 (62.4%) were conventional synthetic DMARDs (csDMARDs), with methotrexate being the most frequent (n=553, 48.45%). One-year persistence rates varied between 55.9% (49.2-62.0%) for tumour necrosis factor inhibitors, and 63.4% (59.6-67.0%) for csDMARDs. Treatment adherence, assessed through medication possession ratio, varied between 71.9% and 90.8%, with ≥80% being the adherence cut-off. Almost half of DMARD initiations were associated with long-term (>6 months), high-dose oral steroid use (~7 mg/day prednisone equivalent).

CONCLUSION

Despite a diverse therapeutic arsenal, there remains a medical need that is not covered by current RA management, which is frequently compensated for by overprescription of steroids.

Imperatorin Inhibits Proliferation, Migration, and Inflammation via Blocking the NF-κB and MAPK Pathways in Rheumatoid Fibroblast-like Synoviocytes

Rheumatoid arthritis (RA) is a chronic joint inflammatory disease associated with the aberrant activation of fibroblast-like synoviocytes (FLSs). Searching for natural compounds that may suppress the activation of FLSs has become a complementary approach for RA treatment. Here, we investigated the effects and mechanisms of imperatorin (IPT) on proliferation, migration, and inflammation in primary cultured arthritic FLSs. We found that IPT significantly suppressed TNFα-induced proliferation and migration of arthritic FLSs, but showed little effect on survival and apoptosis. In addition, IPT treatment significantly reduced the TNFα-induced expression of pro-inflammatory cytokines (IL-1β, TNFα, IL-6, and IL-8) in arthritic FLSs. Further mechanism studies suggested that IPT inhibited the activations of p38 and extracellular signal-regulated kinase (ERK). Also, IPT blocked the nuclear factor of κB (NF-κB) activation by suppressing the phosphorylation and degradation of IκBα, thereby preventing the translocation of p65. Collectively, our results demonstrated that IPT could inhibit the over-activated phenotypes of arthritic FLSs via the mitogen-activated protein kinase (MAPK) (p38 and ERK) and NF-κB pathways leading to the down-regulation of pro-inflammatory cytokines, which might be beneficial to the anti-proliferative and anti-migratory activities of FLS cells. These findings suggest that IPT has the potential to be developed as a novel agent for RA treatment.

Impact of persistence with tumour necrosis factor inhibitors on healthcare resource utilization and costs in chronic inflammatory joint diseases

AIM

To assess persistence with subcutaneous (SC) tumour necrosis factor (TNF) inhibitors as well as the impact of persistence on healthcare resource utilization (HCRU) and costs in patients with chronic inflammatory joint diseases.

METHODS

In this cohort study using population-based French claims data (from 2011 to 2014), we measured persistence with SC TNF inhibitors within 12 months (M0-12) following treatment initiation in treatment-naïve and treatment-experienced users (divided into three cohorts: rheumatoid arthritis [RA], ankylosing spondylitis [AS] and psoriatic arthritis [PsA]). Persistent patients were propensity score matched to nonpersistent patients at M12. The impact of persistence status on HCRU and costs was assessed during M12-24.

RESULTS

Of treatment-naïve (n = 3,804) and treatment-experienced (n = 2,279) users, only 56.1% and 46.8% were persistent at M12, respectively. Nonpersistent patients had more outpatient visits, computerized tomography scans, spine or joint magnetic resonance imaging procedures and disease-related hospitalizations, while persistent patients had more rheumatologist visits. Nonpersistent patients had lower drug costs but higher nondrug-related healthcare and hospitalization costs than persistent patients. In AS and PsA, overall healthcare costs were similar in persistent and nonpersistent patients. In RA, overall healthcare costs were lower in persistent patients (15,753€ vs 17,590€ in treatment-naïve and 17,622€ vs 21,177€ in treatment-experienced).

CONCLUSION

Persistence with SC TNF inhibitors within first 12 months following treatment initiation was low in both treatment-naïve and treatment-experienced patients. Differences were observed in distribution of costs between persistent and nonpersistent patients, showing that nonpersistence with SC TNF inhibitors can lead to increased HCRU and higher costs.

Secondary prevention of acute coronary syndrome with antiplatelet agents in real life: A high-dimensional propensity score matched cohort study in the French National claims database

Users of newly marketed drugs often differ from the patients included in randomized clinical trials, and from patients prescribed similar drugs. Cohorts of such users may be compared using propensity score adjustment, or similar user cohorts may be built using high-dimensional propensity score matching in large population databases. One such database is SNDS, the French nationwide claims and hospitalization database, which covers 99 % of the French population. It has yet been rarely used.

To study the comparative effectiveness and safety in secondary coronary prevention of ticagrelor, compared to clopidogrel or prasugrel, we identified in SNDS patients who were dispensed any of the three antiplatelet agents of interest (± aspirin) within a month after discharge from hospital for acute coronary syndrome (ACS) and followed them one year for recurrence of ACS, stroke, acute bleeding, or death. High-dimensional propensity scores were developed to identify matched cohorts. Drug performances were also compared in the whole population using adjustment on the same parameters. Here we describe the database that was used, and the methods developed for the high-dimensional propensity score matching, resulting in standardized mean differences between the matched populations of less than 2 % for all of the 500+ variables included in the model.

•

This study was done in a newly available large-scale claims database, which may differ from other population databases, by it size and exhaustiveness

•

The methods elaborate on standard high-dimensional propensity scores as adapted to this claims database

Comparative Real-Life Effectiveness and Safety of Dabigatran or Rivaroxaban vs. Vitamin K Antagonists: A High-Dimensional Propensity Score Matched New Users Cohort Study in the French National Healthcare Data System SNDS

BACKGROUND

Clinical trials have indicated that the direct-acting oral anticoagulants dabigatran and rivaroxaban have better risk/benefit profiles than do vitamin K antagonists (VKAs) for stroke prevention in non-valvular atrial fibrillation (NVAF).

OBJECTIVE

Our objective was to compare the 1-year real-life risk of major clinical events with dabigatran or rivaroxaban versus VKAs for NVAF.

METHODS

This was a high-dimensional propensity score (hdPS)-matched cohort study of new users of dabigatran, rivaroxaban or VKAs for NVAF in the French national healthcare systems database in 2013 followed-up for 1 year [22]. Hazard ratios (HRs) with 95% confidence intervals (CIs) for clinical events and death were determined during exposure.

RESULTS

In 2013, a total of 103,101 new anticoagulant users had definite NVAF: 44,653 VKA, 27,060 dabigatran, and 31,388 rivaroxaban. In matched populations, HRs were as follows for dabigatran versus VKAs (20,489 per group): stroke and systemic embolism (SSE) 0.75 (95% CI 0.63-0.88), clinically relevant bleeding (CRB) 0.58 (95% CI 0.51-0.66), hemorrhagic stroke (HS) 0.22 (95% CI 0.14-0.36), gastrointestinal bleeding (GIB) 0.98 (95% CI 0.80-1.19), acute coronary syndrome (ACS) 0.79 (95% CI 0.65-0.95), death 0.74 (95% CI 0.67-0.82), composite (any of the above) 0.71 (95% CI 0.66-0.76). For matched rivaroxaban versus VKA (23,053 per group) HRs were as follows: SSE 0.98 (95% CI 0.85-1.14), CRB 0.83 (95% CI 0.75-0.92), HS 0.65 (95% CI 0.49-0.87), GIB 1.08 (95% CI 0.90-1.30), ACS 0.84 (95% CI 0.71-1.00), death 0.77 (95% CI 0.71-0.84), composite 0.84 (95% CI 0.79-0.89). Numbers needed to treat to observe one fewer death were 49 ± 0.05 with dabigatran or rivaroxaban versus VKAs.

CONCLUSION

Consistent with results from clinical trials and other observational studies, dabigatran and rivaroxaban were at least as effective and safer than VKAs for the prevention of thromboembolic events in NVAF over 1 year in the French population.

STUDY REGISTRATION

European Medicines Agency EUPAS 13017 (www.encepp.eu) Clinicaltrials.gov id NCT02785354.

Comparative effectiveness of direct oral anticoagulants versus low-molecular weight heparins for the prevention of venous thromboembolism after total hip or knee replacement: A nationwide database cohort study

BACKGROUND

Venous thromboembolism (VTE) after total knee or hip replacement (TKR, THR) is usually prevented with low-molecular weight heparin (LMWH), and increasingly by direct oral anticoagulants (DOAC). The aim of the present study was to compare the benefit-risk and medical costs of DOAC vs. LMWH in a real-life setting.

METHODS

All patients with THR or TKR in France between Jan-1st 2013 and Sep-30th 2014, discharged to home, were identified and followed-up for 3 months in the French nationwide claims database, SNDS. DOAC users were 1:1 matched with LWMH users on gender, age and propensity score. Relative risks (RR) of hospitalized VTE, hospitalized bleeding and death were estimated using quasi-Poisson models. Medical costs were calculated according to the societal perspective, including total cost for outpatient claims and national DRG costs for hospitalisations.

RESULTS

Most DOAC users (≥ 98.8%) were matched to a LMWH patient. For the 63,238 matched THR patients, the 3-month absolute risk of VTE was 0.9‰ with DOAC and 2.5‰ with LMWH (RR = 0.35 [0.23 to 0.54]), of bleeding 1.8‰ and 2.1‰ (0.88 [0.62-1.25]), death 0.7‰ and 1.1‰ (0.68 [0.40-1.15]). For the 31,440 matched TKR patients, risks were 1.6‰ and 2.3‰ (0.69 [0.42-1.16]) for VTE, 2.4‰ and 3.8‰ (0.64 [0.43 to 0.97]) for bleeding, and 0.6‰ and 0.8‰ (0.69 [0.30-1.62]) for all-cause death. Mean medical costs were 28% and 21% lower with DOAC than LMWH for THR and TKR, respectively. This nationwide study found a very low risk of VTE, hospitalized bleeding and death after THR or TKR discharge in patients with VTE prevention in real-life setting, with better benefit-risk profiles of DOAC compared to LMWH, and associated cost savings.

Effectiveness and safety of 110 or 150 mg dabigatran vs. vitamin K antagonists in nonvalvular atrial fibrillation

AIMS

We compared the 1-year safety and effectiveness of dabigatran 110 mg (D110) or 150 mg (D150) twice daily to vitamin K antagonists (VKA) in patients with nonvalvular atrial fibrillation.

METHODS

New user cohort study of patients dispensed D110 or D150 vs. VKA in 2013 for nonvalvular atrial fibrillation, followed 1 year in the French Système National des Données de Santé (66 million persons). D110 and D150 users were matched 1:1 with VKA users on sex, age, date of first drug dispensing and high-dimensional propensity score. Hazard ratios [HR (95% confidence intervals)] for stroke and systemic embolism (SSE), major bleeding (MB) and death were computed using Cox proportional hazards or Fine and Gray models during exposure.

RESULTS

In 14 442 matched D110 and VKA patients, mean age 79, 49% male, 91% with CHA₂ DS₂ -VASc ≥2 and 8% with HAS-BLED score >3, incidence rates of SSE were 1.9% and 2.6% person-years [HR 0.69 (0.56-0.84)], MB 1.8% and 2.9% [0.62 (0.51-0.76)], death 7.2% and 8.6% [0.84 (0.76-0.94)]. In 8389 matched D150 and VKA patients, mean age 67, 67% male, 65% with CHA₂ DS₂ -VASC ≥2; < 5% HAS-BLED >3, incidence rates were for SSE 1.4% and 1.9% [0.76 (0.56-1.04)], MB 0.6% and 1.9% [0.30 (0.20-0.46)], death 1.6% and 3.6% [0.46 (0.35-0.59)]. Numbers needed to treat to observe one fewer death were 78 for D110, 88 for D150.

CONCLUSION

In real life D110 and D150 were at least as effective, and safer than VKA.

Predicting Acute Exacerbations in Chronic Obstructive Pulmonary Disease

BACKGROUND

With increasing health care costs that have outpaced those of other industries, payers of health care are moving from a fee-for-service payment model to one in which reimbursement is tied to outcomes. Chronic obstructive pulmonary disease (COPD) is a disease where this payment model has been implemented by some payers, and COPD exacerbations are a quality metric that is used. Under an outcomes-based payment model, it is important for health systems to be able to identify patients at risk for poor outcomes so that they can target interventions to improve outcomes.

OBJECTIVE

To develop and evaluate predictive models that could be used to identify patients at high risk for COPD exacerbations.

METHODS

This study was retrospective and observational and included COPD patients treated with a bronchodilator-based combination therapy. We used health insurance claims data to obtain demographics, enrollment information, comorbidities, medication use, and health care resource utilization for each patient over a 6-month baseline period. Exacerbations were examined over a 6-month outcome period and included inpatient (primary discharge diagnosis for COPD), outpatient, and emergency department (outpatient/emergency department visits with a COPD diagnosis plus an acute prescription for an antibiotic or corticosteroid within 5 days) exacerbations. The cohort was split into training (75%) and validation (25%) sets. Within the training cohort, stepwise logistic regression models were created to evaluate risk of exacerbations based on factors measured during the baseline period. Models were evaluated using sensitivity, specificity, and positive and negative predictive values. The base model included all confounding or effect modifier covariates. Several other models were explored using different sets of observations and variables to determine the best predictive model.

RESULTS

There were 478,772 patients included in the analytic sample, of which 40.5% had exacerbations during the outcome period. Patients with exacerbations had slightly more comorbidities, medication use, and health care resource utilization compared with patients without exacerbations. In the base model, sensitivity was 41.6% and specificity was 85.5%. Positive and negative predictive values were 66.2% and 68.2%, respectively. Other models that were evaluated resulted in similar test characteristics as the base model.

CONCLUSIONS

In this study, we were not able to predict COPD exacerbations with a high level of accuracy using health insurance claims data from COPD patients treated with bronchodilator-based combination therapy. Future studies should be done to explore predictive models for exacerbations.

DISCLOSURES

No outside funding supported this study. Samp is now employed by, and owns stock in, AbbVie. The other authors have nothing to disclose. Study concept and design were contributed by Joo and Pickard, along with the other authors. Samp and Lee performed the data analysis, with assistance from the other authors. Samp wrote the manuscript, which was revised by Schumock and Calip, along with the other authors.

DPP-4 Inhibitor-Induced Rheumatoid Arthritis Among Diabetics: A Nested Case-Control Study

INTRODUCTION

The risk of rheumatoid arthritis (RA) associated with dipeptidyl peptidase-4 inhibitor (DPP-4i) use is unclear. This study assesses the RA risk associated with DPP-4i use among a diabetic cohort initiating second-line therapy.

METHODS

This was a nested case-control study, using the adult diabetic population starting second-line antidiabetic therapy from IMS LifeLink Plus^® database (2006-2015). Cases were those with two or more RA diagnosis, at least one prescription, and 180 days enrollment prior to the event date (earliest of the two: first RA diagnosis, first RA prescription). Controls were drawn from the nest after matching (1:15) with cases on index date (± 90 days), age (± 5 years), sex, and event date (imputed to have the same time difference between cohort entry and event date as the matched case). Exposure and covariate information was gathered from the 180-day period prior to event date. Conditional logistic regression was used to assess exposure among cases and controls. Adjusted analysis was carried out after controlling for important medications and comorbidities.

RESULTS

The final sample consists of 790 cases and 11,850 controls; of these, 151 cases (19.11%) and 2177 controls (18.37%) had DPP-4i claims during the exposure assessment period. DPP-4i therapy was not significantly associated with the development of RA after adjusting for covariates (OR = 1.156, 95% CI 0.936-1.429). Changing the exposure definition or exposure window to 1 year and subgroup analyses yielded similar results except for the non-insulin-using subgroup (OR = 1.299, 95% CI 1.001-1.985) which showed a significant positive association.

CONCLUSION

DPP-4i were not significantly associated with the risk of RA compared with other second-line antidiabetic therapies.

Exchangeable Femoral Neck (Dual-Modular) THA Prostheses Have Poorer Survivorship Than Other Designs: A Nationwide Cohort of 324,108 Patients

BACKGROUND

Exchangeable neck stems, defined as those with a dual taper (that is, a modular junction between the femoral head and the femoral neck and an additional junction between the neck and the stem body), were introduced in THA to improve restoration of joint biomechanics (restoring anteversion, offset, and limb length) and reduce the risk of dislocation. However exchangeable necks have been reported to result in adverse effects such as stem fractures and acute local tissue reaction. Whether they result in a net improvement to or impairment of reconstructive survivorship remains controversial.

QUESTIONS/PURPOSES

(1) To compare the prosthetic survivorship and all-cause revision risk of exchangeable femoral neck THAs versus fixed neck THAs, taking known prosthetic revision risk factors into account; and (2) to compare the cause-specific revision risk of exchangeable femoral neck THAs versus fixed neck THAs, adjusting for known prosthetic risk factors.

METHODS

Using French national health-insurance databases, we identified all French patients older than 40 years who underwent primary THA from 2009 through 2012. To ensure accuracy of the data, we considered only beneficiaries of the general insurance scheme (approximately 77% of the population). Characteristics of the prosthesis and the patients receiving an exchangeable femoral neck THA were compared with those receiving a fixed femoral neck THA (defined as femoral stem with only the head being exchangeable). Revision was the event of interest. Followup started on the date the THA was performed, until the patient experienced revision, died, was lost to followup, or until the followup period ended (December 31, 2014), whichever came first. Competing risk THA survivorship was calculated and compared (purpose 1), as were cause-specific Cox regression models (purpose 2). The study cohort included 324,108 individuals with a mean age of 77 years. A total of 24% underwent THA for acute trauma, and 3% of the group received an exchangeable neck THA. During the median 45-month followup (mean, 42 months; minimum, 1 day; maximum, 6 years), 11,968 individuals underwent prosthetic revision.

RESULTS

The cumulative revision incidence was 6.5% (95% CI, 5.8%-7.3%) for exchangeable neck THAs versus 4.7% (95% CI, 4.6%-4.8%) for fixed neck THAs (p < 0.001). After controlling for potential confounding variables including age, sex, comorbidities, indication for THA, cementation, bearing surface, and the characteristics of the center where the implantation was performed, we found that the exchangeable femoral neck THA was associated with an increased hazard ratio (HR) of revision of 1.26 (95% CI, 1.14-1.38; p < 0.001) compared with the fixed neck THA. When dealing with cause-specific revision, exchangeable neck THAs had a higher incidence of revision for implant failure or periprosthetic fracture, and for mechanical complications; adjusted HRs were, respectively, 1.68 (95% CI, 1.24-2.27; p < 0.001) and 1.27 (95% CI, 1.13-1.43; p < 0.001), for exchangeable neck THAs compared with fixed ones.

CONCLUSIONS

Exchangeable neck THAs had poorer survivorship independent of other prosthetic revision risk factors. Accordingly, expected anatomic and functional benefits should be carefully assessed before choosing this design.

LEVEL OF EVIDENCE

Level III, therapeutic study.

Relative impact of hospital and surgeon procedure volumes on primary total hip arthroplasty revision: a nationwide cohort study in France

BACKGROUND

Both surgeon and hospital procedure volumes have been found to be associated with total hip arthroplasty (THA) outcomes. However, little research has been conducted on the relative influence. We studied the association between THA survivorship and both hospital and surgeon procedure volumes, considering their relative impact.

METHODS

A population-based cohort included all patients aged ≥40 years having received a unilateral primary THA from 2010 to 2011, from the French National Health Insurance Database. Patients were followed up until the end of 2014. The outcome was THA revision. Exposures of interest were procedure volumes, divided into tertiles: <1.5, 1.5-4, >4 and <7, 7-15, >15 procedures per month defined as low, medium, and high volumes for surgeon and hospital, respectively.

RESULTS

The cohort had 62,906 patients, with mean age 69 years and women 57%. Mean surgeon and hospital volumes were 8 and 23 procedures per month, respectively, and 5%, 72%, 22% and 7%, 28%, 65% of THAs were implanted by a low-, medium-, and high-volume surgeon or in a low-, medium-, and high-volume hospital, respectively. Median follow-up was 45 months (range, 0-57 months). In multivariate analysis, adjusted for both surgeon and hospital volumes, for patient and THA characteristics, a lower surgeon volume was associated with poorer THA survivorship (adjusted hazard ratio [aHR] = 1.19; 95% confidence interval [CI], 1.07-1.34 and aHR = 1.70; 95% CI, 1.40-2.05, for medium- and low-volume surgeon, respectively, compared with that of high volume), whereas hospital volume was not.

CONCLUSIONS

This study brings evidence to support the notion that THAs performed by high-volume surgeons in French private hospitals have higher survivorship in the first 4 years.

Outcomes in patients after myocardial infarction similar to those of the PEGASUS-TIMI 54 trial: A cohort study in the French national claims database

AIMS

The present study aims to describe real-life outcomes in stable patients after-myocardial infarction (MI) similar to those in the PEGASUS-TIMI 54 trial (PEGASUS), which found long-term benefits of ticagrelor in patients with a history of MI.

METHODS

One-year event-free post-MI patients were identified in the French claims database representative 1/97 sample (2005-2010) and followed for up to 3 years. A PEGASUS-like (PL) population included patients with age ≥ 65 years, or age ≥ 50 and diabetes, renal dysfunction or prior MI, without stroke, end-stage renal failure or oral anticoagulation. Outcomes were: a composite of all-cause death or hospital admission for MI or stroke; individual events; major bleeding.

RESULTS

There were 1585 post-MI patients totalling 3926 person-years including 865 PL patients (2114 PY); 68% were male; mean age was 66 (standard deviation 15) in post-MI, 74 (10) in PL. Outcomes per 100 person-years [95% confidence interval] were, respectively, in post-MI and PL 6.3 [5.6-7.1] and 7.8 [6.7-8.9] for the composite outcome; 5.1 [4.4-5.8] and 6.5 [5.5-7.6] for death; 1.0 [0.7-1.3] and 1.0 [0.6-1.4] for MI; 0.6 [0.4-0.9] and 0.9 [0.5-1.2] for stroke; 1.3 [0.9-1.6] and 1.4 [0.9-1.9] for major bleeding. Event rates were stable over the 3 study years. Placebo patients in the PEGASUS-TIMI54 Study were younger, more often male and had lower event rates, especially for all-cause death and major bleeding.

CONCLUSIONS

Patients selected using the criteria described in PEGASUS were older with more comorbidities, resulting in higher all-cause death and bleeding rates, but similar MI recurrence rates.

Sources of European drug consumption data at a country level

OBJECTIVES

This study aimed at outlining the characteristics of nationwide administrative databases monitoring drug consumption in Europe.

METHODS

Internet and bibliographic databases (April 2010) were searched and experts in drug utilization (DU) research interviewed to find nationwide administrative medicines consumption databases in Europe, with data for the out- and inpatient healthcare sector. A questionnaire was developed to gather additional information. We collected data providers, websites, accessibility, data sources, healthcare settings, population coverage, medicines-related data, patient and prescriber data, periods covered, and linkage to other databases.

RESULTS

Thirty-one administrative nationwide medicine consumption databases in 25 countries were identified. Questionnaires were responded for 20 databases. Eleven provided wholesalers' sales data, 11 on reimbursed, 5 on prescribed, and 4 on dispensing medicines. Fifteen databases provided inpatient drug consumption data, mainly wholesalers' sales.

CONCLUSIONS

Nationwide administrative databases are of value to all stakeholders involved in the conduct and interpretation of post-marketing safety studies, and in the conduct of DU research. The endorsement of the anatomical therapeutic chemical/defined daily dose methodology by these databases contributes to data harmonization. However, there is still a lack of information on inpatient medicines consumption at a patient-level.

Self-controlled case series and misclassification bias induced by case selection from administrative hospital databases: application to febrile convulsions in pediatric vaccine pharmacoepidemiology

Vaccine safety studies are increasingly conducted by using administrative health databases and self-controlled case series designs that are based on cases only. Often, several criteria are available to define the cases, which may yield different positive predictive values, as well as different sensitivities, and therefore different numbers of selected cases. The question then arises as to which is the best case definition. This article proposes new methodology to guide this choice based on the bias of the relative incidence and the power of the test. We apply this methodology in a validation study of 4 nested algorithms for identifying febrile convulsions from the administrative databases of 10 French hospitals. We used a sample of 695 children aged 1 month to 3 years who were hospitalized in 2008-2009 with at least 1 diagnosis code of febrile convulsions. The positive predictive values of the algorithms ranged from 81% to 98%, and their sensitivities were estimated to be 47%-99% in data from 1 large hospital. When applying our proposed methods, the algorithm we selected used a restricted diagnosis code and position on the discharge abstract. These criteria, which resulted in the selection of 502 cases with a positive predictive value of 95%, provided the best compromise between high power and low relative bias.