Physician Claims Research Articles

Tramadol and the risk of seizure: nested case-control study of US patients with employer-sponsored health benefits

ObjectivesTramadol is a widely prescribed analgesic that influences both opioid and monoamine neurotransmission. While seizures have been reported with its use, the risk in clinical practice has not been well...

Association between strained ICU capacity and healthcare costs in Canada: A population-based cohort study

BackgroundIntensive care is resource intensive, with costs representing a substantial quantity of total hospitalization costs. Strained ICU capacity compromises care quality and adversely impacts outcomes; however, the association between strain and healthcare costs has not been explored. Materials and methodsPopulation-based cohort study performed in 17 adult ICUs in Alberta, Canada. Data were captured on hospitalizations, ambulatory care, physician services and drug dispenses occurring 1-year before and 1-year after index ICU admission. Strain was defined as occupancy ≥90%; with 21 additional definitions evaluated. Patients were categorized as strain and non-strain admissions. Costs attributable to strain, were calculated as difference-in-difference costs using propensity-score matching. Results30,557 patients were included (strain: 11,830 [38.7%]; non-strain: 18,727 [61.3%]). At 1-year, strain admissions had adjusted-incremental per-patient cost of CA$9406 (95%CI, $5654-13,157) compared to non-strain admissions, due to hospitalization costs (CA$7930; 95%CI, $4553-11,307) and physician claims (CA$844; 95%CI, $430–1259). This equated to CA$111.3 million (95%CI, $66.9–155.6 million) in excess attributable costs. Strain portended longer hospitalization (3.3 days; 95%CI, 1.1–5.5); and more ambulatory visits (1.0; 95%CI, 0.1–2.0) and physician claims (9.5; 95%CI, 6.2–12.7). Incremental costs were robust across strain definitions. ConclusionsAdmissions to ICUs experiencing strain incur incremental costs, attributed to longer hospitalization and physician services.

Burden and Cost of Outpatient Hemorrhoids in the United States Employer-Insured Population, 2014

Although hemorrhoids are a common indication for seeking health care, there are no contemporary estimates of burden and cost. We examined data from an administrative claims database to estimate health care use and aggregate costs. We conducted a cross-sectional study using the MarketScan Commercial Claims and Encounters Database for 2014. The analysis included 18.9 million individuals who were aged 18-64 and continuously enrolled with prescription coverage. Outpatient hemorrhoid claims were captured using the International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes in the first position, as well as Common Procedural Terminology codes. Prescription medications were identified using National Drug Codes. Annual prevalence and costs were determined by summing gross payments for prescription medications, physician encounters, and facility costs. We used validated weights to standardize annual cost estimates to the US employer-insured population. In 2014, we identified 227,638 individuals with at least one outpatient hemorrhoid-related claim (annual prevalence, 1.2%). Among those, 119,120 had prescription medication claims, 136,125 had physician claims, and 28,663 had facility claims. After standardizing, we estimated that 1.4 million individuals in the US employer-insured population sought care for hemorrhoids in 2014 for a total annual cost of $770 million. This included $322 million in physician claims, $361 million in outpatient facility claims, and $88 million in prescription medication claims. The estimated economic burden of hemorrhoids in the employer-insured population approaches $800 million annually. Given the substantial and rising burden and cost, expanded research attention should be directed to hemorrhoidal etiology, prevention, and treatment.

Validation of a case definition for depression in administrative data against primary chart data as a reference standard

BackgroundBecause the collection of mental health information through interviews is expensive and time consuming, interest in using population-based administrative health data to conduct research on depression has increased. However, there is concern that misclassification of disease diagnosis in the underlying data might bias the results. Our objective was to determine the validity of International Classification of Disease (ICD)-9 and ICD-10 administrative health data case definitions for depression using review of family physician (FP) charts as the reference standard.MethodsTrained chart reviewers reviewed 3362 randomly selected charts from years 2001 and 2004 at 64 FP clinics in Alberta (AB) and British Columbia (BC), Canada. Depression was defined as presence of either: 1) documentation of major depressive episode, or 2) documentation of specific antidepressant medication prescription plus recorded depressed mood. The charts were linked to administrative data (hospital discharge abstracts and physician claims data) using personal health numbers. Validity indices were estimated for six administrative data definitions of depression using three years of administrative data.ResultsDepression prevalence by chart review was 15.9–19.2% depending on year, region, and province. An ICD administrative data definition of ‘2 depression claims with depression ICD codes within a one-year window OR 1 discharge abstract data (DAD) depression diagnosis’ had the highest overall validity, with estimates being 61.4% for sensitivity, 94.3% for specificity, 69.7% for positive predictive value, and 92.0% for negative predictive value. Stratification of the validity parameters for this case definition showed that sensitivity was fairly consistent across groups, however the positive predictive value was significantly higher in 2004 data compared to 2001 data (78.8 and 59.6%, respectively), and in AB data compared to BC data (79.8 and 61.7%, respectively).ConclusionsSensitivity of the case definition is often moderate, and specificity is often high, possibly due to undercoding of depression. Limitations to this study include the use of FP charts data as the reference standard, given the potential for missed or incorrect depression diagnoses. These results suggest that that administrative data can be used as a source of information for both research and surveillance purposes, while remaining aware of these limitations.

Incidence of Osteoporotic Refractures Following Proximal Humerus Fractures in Adults Aged 50 Years and Older in Korea.

BackgroundThe purpose of this study was to investigate the incidence and characteristics of osteoporotic refractures after proximal humerus fracture in Korean adults aged above 50 years.MethodsPatients aged 50 years or older with initial proximal humerus fractures reported in 2012 were enrolled and followed up until 2016 using the Korean National Health Insurance data. Based on the last claim date, the refractures were classified as osteoporotic fractures including spine, hip, distal radius, and humerus 6 months after the index fracture involving the proximal humerus. Each osteoprotic fracture was identified using specific International Classification of Diseases, 10th Revision codes and site-specific physician claims for procedures.ResultsA total of 5,587 first-time fractures involving proximal humerus were reported in 2012. Among them, a total of 1,018 osteoporotic refractures occured between 2012 and 2016. The total cumulative incidence of osteoporotic refractures was 4.85% (271/5,587) at 1 year, 9.61% (537/5,587) at 2 years, 14.21% (794/5,587) at 3 years, and 18.22% (1,018/5,587) at 4 years. In terms of site by year, the incidence of associated refractures was as follows: spine, 48.62% (495/1,018); hip, 25.83% (263/1,018); wrist 18.57% (189/1,018); and humerus 6.97% (71/1,018) during all the follow-up periods.ConclusionsOur study showed that the cumulative incidence of osteoporotic refractures following proximal humerus fractures in the elderly population has been increasing over the years. Given that osteoporotic refractures are associated with an increased mortality risk, a public health strategy to prevent the refracture after proximal humerus fracture in the elderly is imperative.

Increasing Incidence and Prevalence of Pathologic Hemoglobinopathies Among Children in Ontario, Canada from 1991-2013

Abstract Background: In Ontario, Canada's largest province, population-based health administrative data represents an accessible and useful tool for population surveillance of people with chronic diseases. While hemoglobinopathies can be identified using data from universal hemoglobinopathy screening, which was implemented in November 2006, these data would not contain information on affected immigrants (21.9% of the population). We validated algorithms using provincial health administrative data and newborn screening data to identify children with hemoglobinopathies whether or not they were born in Ontario, thereby creating a population-based surveillance cohort. Objectives: (1) Validate algorithms to identify children with sickle cell disease, thalassemia and other hemoglobinopathies from within health administrative data; and (2) Determine incidence and prevalence of hemoglobinopathies in Ontario children. Methods: For the validation study, a positive reference cohort was established using lists of known hemoglobinopathy patients who were followed at five pediatric hemoglobinopathy treatment sites in Ontario and born between November 24, 2006 and March 31, 2013. Health card numbers of these patients were linked deterministically to unique identification numbers in administrative data, which included data on hospitalizations, physician claims, sociodemographic characteristics, immigration records and cause of death. The negative reference cohort included all children residing in Ontario cities who had never been seen at a hemoglobinopathy centre, and therefore assumed not to have disease. Various combinations of administrative data codes were tested for their ability to identify children <18 years of age with hemoglobinopathies from within the databases, and we selected the algorithms with the highest positive predictive value, while maintaining sensitivity >80%. Using two validated algorithms, we identified all children with hemoglobinopathies born between April 1, 1991 and March 31, 2013. We described the crude incidence and prevalence per 100,000 patient-years (PYs). Results: Two algorithms functioned best to identify incident and prevalent hemoglobinopathy cases (see Table). Among children born between April 1, 1991 to March 31, 2013, 1526 incident hemoglobinopathy patients were identified using Algorithm 1 (crude incidence of 4.85 per 100,000 PYs) and 1660 new hemoglobinopathy patients were identified using Algorithm 2 (crude incidence 5.28 per 100,000 PYs, 95% CI 3.51 to 3.92). In 2013, the overall prevalence of children <18 years living with hemoglobinopathies was 1215-1325 cases. Conclusion: Through an innovative approach using provincial health administrative, immigration and demographic data, this study identified a rising incidence and prevalence of hemoglobinopathies among Ontario children <18 years of age between April 1, 1991 and March 31, 2013, potentially due to increased immigration rates. These findings could be used to inform health services distribution. This surveillance cohort will be used to understand the impact of immigration status on health care inequality for hemoglobinopathy-related health services delivery, as well as to assess outcomes in this important group of chronic diseases. Disclosures Klaassen: Amgen Inc.: Membership on an entity's Board of Directors or advisory committees; Octapharma AG: Consultancy, Honoraria; Agios Pharmaceuticals Inc.: Consultancy; Novartis: Research Funding; Hoffman-La Roche: Consultancy; Shire: Consultancy; Cangene: Research Funding. Jardine:Pfizer: Other: Advisory board; Bayer: Other: Advisory board; Baxalta: Other: Advisory board.

2514. Pediatric Medically Attended Shingles in Alberta, Canada 2016; Preliminary Results

BackgroundHerpes zoster (shingles) is a reactivation of latent varicella virus. Shingles is uncommon in children and children vaccinated against varicella may be at a reduced risk of shingles. Alberta has a publicly funded healthcare system and added publicly funded varicella vaccine to the routine childhood vaccination schedule in 2001.MethodsWe used provincially held administrative health databases to examine the epidemiology of incident shingles cases in children under the age of 19. Incident shingles cases were defined as the earliest record of ICD-9-CM 053 OR ICD-10-CA B02 coded physician claims, hospital, or emergency room visits between 1985 and 2016, with incident cases in this cohort occurring between January 1, 2016 and December 31, 2016. Varicella immunization was identified through Alberta’s immunization repository and immunosuppressive comorbid conditions (neoplasms, HIV/AIDS, cystic fibrosis, and immune system disorders) were identified using ICD diagnostic codes from physician claims, hospital, or emergency room visits and Alberta’s Communicable Disease Control databases.Results1,003 incident shingles cases were identified in children under the age of 19 in 2016, a crude rate of 0.98/1,000 persons. Females comprised 54% of cases. The largest proportion of cases occurred among those aged 15–19 years. About 39% of cases were prescribed antiviral medication, most commonly those aged 15–19 years. The crude rate per 1,000 population increased with age: 0.5 for children under the age of 1, 1.2 for those 1–4 years, 1.25 for children 5–9 years, 2.19 for children 10–14 years, and 3.7 for children aged 15–19 years. Crude rates were similar among both males and females. Less than 3% of the cases had ever been immunized against varicella. Shingles diagnostic codes were not validated, which likely led to an overestimation of the true rates of disease.ConclusionAdditional studies are needed on pediatric shingles cases and factors that influence shingles in this group, as well as validation studies of ICD diagnostic coding in administrative data.Disclosures M. L. Russell, Novartis Pharmaceuticals Canada Inc.: Grant Investigator and Unconditional Research Grant, Grant recipient. Merck Frosst Canada Inc.: Grant Investigator and Unconditional Research Grant, Grant recipient.

122. All-Cause Mortality Increased With Nontuberculous Mycobacterial Lung Disease in US Medicare

BackgroundNontuberculous Mycobacterial Lung Disease (NTMLD) is a chronic, debilitating, and progressive disease. This study evaluates all-cause mortality in patients with NTMLD in the US Medicare.MethodsPatients (n = 43,394) were identified from the Medicare database (excluding Part C) based on physician claims for NTMLD on ≥2 separate occasions ≥30 days apart between 2007 and 2015. About 12% patients were <65 years and qualified for Medicare due to disability. A control cohort (n = 84,814) was randomly selected and matched to the NTMLD sample by age and sex. The NTMLD diagnosis date was assigned to the matched controls as an index date. Poisson and Cox regression were used to derive descriptive rates and adjusted risk of mortality accounting for baseline comorbidities of pulmonary, immune, cardiovascular, cancer, and other disorders.ResultsMean age was 74 (±10) years and 68% were female in both NTMLD and control cohorts. Mean Charlson comorbidity index (CCI) was 2.9 (standard deviation ±2.6) in NTMLD vs. 1.3 (±1.9) in control cohort. In Medicare members ≥65 years, mean age was 76 (±7) years and 70% were female. Mean CCI was 2.8 (±2.5) in NTMLD cohort vs. 1.4 (±2.0) in control cohort. In Medicare members <65, mean age was 53 (±10) and 49% were female. Mean CCI was 3.8 (±3.3) in NTMLD vs. 1.1 (±1.9) in the control. Observed yearly mortality rates were 9.8% in NTMLD vs. 4.7% in control cohort (rate ratio [RR] = 2.1; 95% CI: 2.03–2.13). In ≥65 Medicare members, the observed rates were 9.7% in NTMLD vs. 5.0% in control cohort (RR = 2.0; 1.9–2.0). In Medicare members <65, the observed rates were 10.4% in NTMLD vs. 2.5% in control cohort (RR = 4.1; 3.8–4.5). Compared with the Asian race, observed mortality was higher in NTMLD patients of Native American (hazard ratio [HR] = 1.69, 1.30–2.19), Black (HR = 1.23; 1.08–1.39), Hispanic (HR = 1.27, 1.07–1.51), or White (HR = 1.18, 1.06–1.31) race (Figure 1). Mortality rates were elevated with NTMLD relative to controls in all age categories from ≥65 years (Figure 2). Adjusted mortality increased with NTMLD by 35% overall (HR = 1.35; 1.3–1.4), by 23% in age group ≥65 (HR = 1.23, 1.19–1.27), and almost doubled in age group <65 (HR = 1.97, 1.80–2.15).ConclusionAmong US Medicare enrollees, NTMLD was associated with a 35% increased risk of mortality overall.Figure 1:Figure 1: Observed mortality by race groupsFigure 2:Observed mortality by age groupsDisclosures T. Marras, Insmed Incorporated: Investigator, Consulting fee and Research grant. Horizon Pharmaceuticals: Consultant, Consulting fee. Red Hill: Consultant, Consulting fee. AstraZeneca: CME, Speaker honorarium. Q. Zhang, Insmed Incorporated: Employee, Salary. G. Eagle, Insmed Incorporated: Employee, Salary. P. Wang, Insmed Incorporated: Employee, Salary. R. Zhang, Insmed Incorporated: Consultant, Consulting fee. E. Chou, Insmed Incorporated: Employee, Salary. K. L. Winthrop, Insmed Incorporated: Consultant and Scientific Advisor, Consulting fee and Research grant.

Accuracy of health administrative data to identify cases of reportable travel or migration-related infectious diseases in Ontario, Canada.

An ongoing challenge of estimating the burden of infectious diseases known to disproportionately affect migrants (e.g. malaria, enteric fever) is that many health information systems, including reportable disease surveillance systems, do not systematically collect data on migrant status and related factors. We explored whether health administrative data linked to immigration records offered a viable alternative for accurately identifying cases of hepatitis A, malaria and enteric fever in Ontario, Canada. Using linked health care databases generated by Ontario’s universal health care program, we constructed a cohort of medically-attended individuals with presumed hepatitis A, malaria or enteric fever in Peel region using diagnostic codes. Immigrant status was ascertained using linked immigration data. The sensitivity and positive predictive value (PPV) of diagnostic codes was evaluated through probabilistic linkage of the cohort to Ontario’s reportable disease surveillance system (iPHIS) as the reference standard. Linkage was successful in 90.0% (289/321) of iPHIS cases. While sensitivity was high for hepatitis A and enteric fever (85.8% and 83.7%) and moderate for malaria (69.0%), PPV was poor for all diseases (0.3–41.3%). The accuracy of diagnostic codes did not vary by immigrant status. A dated coding system for outpatient physician claims and exclusion of new immigrants not yet eligible for health care were key challenges to using health administrative data to identify cases. Despite this, we show that linkages of health administrative and immigration records with reportable disease surveillance data are feasible and have the potential to bridge important gaps in estimating burden using either data source independently.

Relative risk of all-cause mortality in patients with nontuberculous mycobacterial lung disease in a US managed care population.

The risk of all-cause mortality of nontuberculous mycobacterial lung disease (NTMLD) in the United States (US) population is not well established. This study aims to assess the public health burden of NTMLD in the US by comparing the relative risk of all-cause mortality in the NTMLD population with an age- and sex-matched cohort from the general population. Patients with physician claims for NTMLD (ICD-9 0.031; ICD-10 A31.0) were identified between 2007 and 2016 from a large US national managed care insurance plan covering approximately 15-18 million members annually. A control group with no NTMLD ICD-9 or 10 codes was randomly selected from the general population and matched 3:1 to the NTMLD sample according to birth year, gender, and insurance benefit coverage. The date of first NTMLD diagnosis of each patient was assigned to the matched controls as the index date. The Cox proportional hazard method compared survival between cohorts, adjusting for demographic factors and baseline comorbidities. A total of 2005 patients with NTMLD and 6014 controls were identified, with a mean follow-up duration of 3.4 years and 3.7 years, respectively. The NTMLD group had substantially higher proportions of patients with asthma (23.3% versus 3.5%), bronchiectasis (36.5% versus 0.1%), COPD (52.0% versus 5.9%), arrhythmia (22.6% versus 6.5%), coronary artery disease (18.5% versus 6.6%), heart failure (11.9% versus 4.1%), and cancer (18.5% versus 5.0%). The unadjusted rate of all-cause mortality from the index date was 20.7 per 1000 person-years in the NTMLD group vs 5.6 per 1000 person-years in the control group (rate ratio = 3.73; 95% CI: 2.93-4.75). Multivariable Cox regression, adjusted for the above variables as well as all other important baseline covariates, showed a doubling risk of all-cause mortality (hazard ratio [HR] = 2.06; CI: 1.52-2.79; P < 0.001) in the NTMLD vs control group. All-cause mortality, adjusted for other factors, more than doubled with NTMLD compared with an age-sex-matched control group in a large US national managed care insurance plan.

Changing patterns of adult asthma incidence: results from the National Health Insurance Service\u2013National Sample Cohort (NHIS-NSC) database in Korea

This study was conducted to assess the changes in the annual incidence of adult asthma in Korea where the prevalence of asthma had increased steadily in recent decades. A population-based cohort study was conducted using the National Health Insurance Service–National Sample Cohort (NHIS-NSC), which consisted of 746,816 adults aged >20 years between 2004 and 2012. Asthma was defined by two or more physician claims on the basis of a primary diagnostic code for asthma and administration of asthma medications within 1 year. The incidence rates and annual percent change were calculated, and the influence of age and sex on the incidence rates was studied. The annual asthma incidence increased from 3.63 in 2004 to 6.07 per 1,000 person-years in 2008. Since 2008, the asthma incidence did not change significantly. The asthma incidence was higher in women than in men throughout the study periods (p < 0.001) and higher in older than younger age groups (p < 0.001). The asthma incidence did not change in all ages since 2008, except for the 20 s who showed a steady increase. The incidence of asthma in adults reached plateau in Korea, which is consistent with the results from studies in other countries.

Breast cancer detection method, diagnostic interval and use of specialized diagnostic assessment units across Ontario, Canada

Breast cancer is detected through screening or through signs and symptoms. In Canada, mammograms for breast cancer screening are offered in organized programs or independently (opportunistic screening). Province of Ontario breast Diagnostic Assessment Units (DAUs) are facility-based programs that provide coordinated breast cancer diagnostic services, as opposed to usual care, in which the primary care provider arranges the tests and consultations. This study describes breast cancer detection method, diagnostic interval and DAU use across Ontario. The study cohort consisted of 6898 women with invasive breast cancer diagnosed in 2011. We used the Ontario Cancer Registry linked to administrative health care databases. We determined the detection method using the Ontario Breast Screening Program (OBSP) data and physician claims. The diagnostic interval was the time between the initial screen, specialist referral or first diagnostic test and the cancer diagnosis. The diagnostic route (whether through DAU or usual care) was determined based on the OBSP records and biopsy or surgery location. We mapped the diagnostic interval and DAU coverage geographically by women's residence. In 2011, 36% of Ontario breast cancer patients were screen-detected, with a 48% rate among those aged 50 to 69. The provincial median diagnostic interval was 32 days, with county medians ranging from 15 to 65 days. Provincially, 48.4% were diagnosed at a DAU, and this ranged from zero to 100% across counties. The screening detection rate in age-eligible breast cancer patients was lower than published population-wide screening rates. Geographic mapping of the diagnostic interval and DAU use reveals regional variations in cancer diagnostic care that need to be addressed.

PSY40 - ALL-CAUSE MORTALITY INCREASED WITH NONTUBERCULOUS MYCOBACTERIAL LUNG DISEASE IN US MEDICARE

Nontuberculous Mycobacterial Lung Disease (NTMLD) is a chronic, debilitating, and progressive disease. This study evaluates all-cause mortality in patients with NTMLD in the US Medicare. Patients (n=43,394) were identified from the Medicare database (excluding Part C) based on physician claims for NTMLD on ≥2 separate occasions ≥30 days apart between 2007 and 2015. About 12% patients were <65 years and qualified for Medicare due to disability. A control cohort (n=84,814) was randomly selected and matched to the NTMLD sample by age and sex. Cox regression was used to derive adjusted risk of mortality accounting for baseline comorbidities of pulmonary, immune, cardiovascular, cancer and other disorders. Mean age was 74 (standard deviation: ±10) years and 68% were female. Mean Charlson comorbidity index (CCI) was 2.9 (±2.6) in NTMLD vs 1.3 (±1.9) in control cohort. Mean CCI was 2.8 (±2.5) in NTMLD vs 1.4 (±2.0) in controls for Medicare members ≥65 years, and 3.8 (±3.3) in NTMLD vs 1.1 (±1.9) in controls for those <65 years. Observed yearly mortality in NTMLD vs controls was 9.7% vs 5.0% respectively among people ≥65 years old (RR = 2.0; 1.9-2.0), and 10.4% vs 2.5% respectively among people <65 years old (RR = 4.1; 3.8-4.5). Adjusted mortality increased with NTMLD by 35% overall (HR = 1.35; 1.3-1.4), by 23% in age group ≥65 (HR = 1.23, 1.19-1.27) and almost doubled in age group <65 (HR = 1.97, 1.80-2.15). NTMLD was associated with a 35% increased risk of all-cause mortality in the overall US Medicare population. Mortality risk almost doubled with NTMLD in the younger member population with disability (<65 years), and increased with NTMLD by 23% in the majority of Medicare population (≥65 years), compared to that in the controls.

Incidence and prevalence of MS in children: A population-based study in Ontario, Canada.

To validate a case definition of multiple sclerosis (MS) in the pediatric population using administrative (health claims) data, and to estimate the incidence and prevalence of MS in the pediatric population for Ontario, Canada. We used population-based administrative data to identify persons aged ≤18 years with MS. We assessed the performance of multiple administrative case definitions using a clinical reference cohort including children with MS, children with monophasic demyelinating syndromes, and healthy children; we report sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). We applied 2 preferred case definitions to estimate the incidence and prevalence of MS from 2003 to 2014. The Canadian Chronic Disease Surveillance System definition of ≥1 hospitalization or ≥5 physician claims for MS within 2 years had a sensitivity of 81.1%, specificity of 100%, PPV of 100%, and NPV of 86%. The Marrie definition of ≥3 hospital or physician claims for MS ever had a sensitivity of 89.2%, specificity of 100%, PPV of 100%, and NPV of 91.5%. Depending on the administrative case definition used, in 2014, the annual age-standardized annual incidence of MS in the pediatric population ranged from 0.99 to 1.24 per 100,000 population, and the age-standardized prevalence ranged from 4.03 to 6.8 per 100,000 population. The prevalence of MS rose over time. Administrative data provide a feasible, valid means of estimating the incidence and prevalence of MS in the pediatric population. MS prevalence in the Ontario pediatric population is among the highest reported in pediatric populations worldwide.

Adherence to and Persistence with Antidepressant Medication during Pregnancy: Does It Differ by the Class of Antidepressant Medication Prescribed?

Pregnant women are often concerned about the impact of medication use on their pregnancy, such as congenital abnormalities. This study examined the rate of adherence to and persistence with antidepressant medications during pregnancy based on the class of antidepressants prescribed. Women who gave birth between 2012 and 2015 in Alberta, Canada; had ≥1 diagnosis of depression within 1 year of preconception in outpatient physician claims, emergency department, or hospitalization administrative data; and were adherent (medication possession ratio ≥80%) to ≥2 consecutive antidepressant prescriptions during the preconception year ( n = 1865) were included in this retrospective cohort study. The rates of adherence and persistence (prescription refill gap ≤30 days) were calculated by antidepressant class and were compared using chi-square tests. During pregnancy, 834 (44.7%; 95% CI, 42.4% to 47.0%) women discontinued antidepressants. Among those continuing antidepressants, the overall rate of adherence was 62.6% (95% CI, 59.4% to 65.7%). The rate differed significantly by medication class ( P < 0.0001), with a rate of 75.1% (95% CI, 68.3% to 80.9%) for serotonin-norepinephrine inhibitors, 60.9% (95% CI, 57.2% to 64.5%) for selective serotonin reuptake inhibitors, 42.8% (95% CI, 19.9% to 69.3%) for nonselective monoamine reuptake inhibitors, and 37.5% (95% CI, 22.5% to 55.4%) for atypical antidepressants. Only, 40.7% (95% CI, 37.5 to 44.1) of women were persistent with antidepressants for the full pregnancy period-the rate differed significantly by medication class ( P < 0.0001). Adherence to and persistence with antidepressants is low during pregnancy and varies by medication class. Low adherence and persistence can interfere with a therapeutic effect of antidepressants, which may contribute to the worsening of depression symptoms.

HEALTHCARE RESOURCE UTILIZATION AMONG CHILDREN WITH CONGENITAL HEART DISEASE: A POPULATION-BASED STUDY

Data regarding healthcare resource utilization (HRU) in early childhood among children with congenital heart disease (CHD) are scarce. Therefore, we sought to describe the extent of HRU incurred among children with CHD in the first five years of life and examine the extent to which it is modulated by the severity of the disease. In this population-based retrospective cohort study, all children born alive between January 2005-March 2014 in Alberta, Canada were included and followed from birth until five years of age. We linked inpatient, outpatient, practitioner claims, and drug dispensing databases with vital statistics (birth and death registries). Children were grouped by CHD severity level as follows: single ventricle (SV), moderate-complex, simple, and non-CHD. Primary outcomes were the number and rate of health service encounters. HRU rate was compared across CHD groups using a univariate Poisson regression model. Among 448,527 children, 223 (0.1%) children had SV lesions, 2,094 (0.5%) had moderate-complex CHD, and 4,690 (1.1%) simple CHD. In the first year of life, the cumulative hospitalization rate/100 children was 335 (95% CI: 312-360) for SV children, 200 (194-206) for moderate-complex CHD, 152 (149-156) for simple CHD vs. 109 (108-109) among non-CHD children (p<0.001). The ambulatory-care visit rate/100 children was 4,871 (4,780-4,963) for SV, 2,278 (2,258-2,299) for moderate-complex, 1,416 (1,405-1,426) for simple CHD vs. 246 (246-247) for non-CHD children (p<0.001). The rates of physician claims and drug-dispensing also demonstrated similar patterns. The median total hospitalization length of stay during the first year of life was 54 days (IQR: 26-95) in SV, 15 (4-39) in moderate-complex, and 6 (2-26) in simple CHD compared with 2 (1-3) among non-CHD children (p<0.001). These differences remained throughout the first five years of life, with CHD children having consistently higher hospitalization rates and ED visit rates in every year of age compared to non-CHD children. In the first year of life, median (IQR) cost per child, excluding drug costs, was $224,324 ($127,596-$312,077) in the SV group, $51,514 ($12,992-$123,332) in moderate-complex CHD, $16,274 ($6,921-$60,069) in simple CHD, and $2,918 ($1,937-$5,480) for non-CHD children (p<0.001). Hospitalizations accounted for the majority of total cost (79.2-84.4% in CHD children vs. 56.4% in non-CHD children). Cumulative HRU is high among CHD children in the first five years of life and increases with increasing CHD severity. As survival rates of children with SV lesions improve, health care systems will require increasing resource allocation to this group.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Point of Care (POC) Influenza Immunization for Pregnant Women, Calgary Zone, Alberta Health Services

IntroductionVaccinating pregnant patients for neonatal protection needs to be integrated into prenatal care as new vaccines emerge. Uptake of influenza vaccine, universally recommended in pregnancy, is low. Immunization was offered and administered to pregnant women at point of care (POC) during two flu seasons at an urban tertiary care center.&#x0D; Objectives and ApproachPrimary objective is to determine if POC impacts immunization rate during flu season among a cohort of pregnant women by location and gestational age. Secondary objectives are to examine the pattern of influenza-like illnesses (ILI) among vaccinated and unvaccinated women, and to describe pilot outcomes of POC. Four consecutive influenza seasons (2014/2015, 2015/2016, 2016/2017, 2017/2018) will be examined using seven databases: a) Clinibase, b) National Ambulatory Care Reporting System; C) Discharge Abstract Database; d) Physician Claims; e) Alberta Perinatal Health Program; f) Calgary Zone Public Health; and g) Pharmaceutical Information Network. Outcomes will be examined descriptively using frequencies and proportions.&#x0D; ResultsBased on the preliminary analysis, approximately 10, 000 visits among 2,500 women occurred during each flu season at the four obstetric care locations: two outpatient clinics and two inpatient units. The proportion of pregnant women who received the flu vaccine ranged from 15-21% during the first three flu seasons. Majority of the women received the vaccine at the flu campaigns (range 48-67%), followed by pharmacy (20-32%). For the 2017-2018 season, year to date uptake rates in outpatient clinics are significantly higher. Final results on additional outcomes will be available by September 2018.&#x0D; Conclusion/ImplicationsIn completing this study, we hope to better understand the patterns of immunization uptake in pregnancy by place of immunization and gestational age, i.e. identifying optimal “window of opportunity”. Results will inform the infrastructure needed to collect data on vaccines administered during pregnancy and linkage to maternal and infant outcomes.

Identifying Knowledge Gaps with Administrative Health Data: A Cohort Study of Traumatic and Non-Traumatic Spinal Cord Injury in Alberta

IntroductionThe Spinal Cord Injury (SCI) population consists of two main sub-groups: traumatic (TSCI) and non-traumatic (NTSCI). TSCI has been studied; however less attention has been given to NTSCI. It is important to understand both SCI sub-groups for identification of knowledge gaps and subsequent health service planning.&#x0D; Objectives and ApproachThe goal is to study the SCI population (both TSCI and NTSCI) in Alberta, Canada, leveraging recent administrative health data. It is difficult to identify NTSCI patients for their heterogeneous conditions, and relatively low prevalence. Consequently, we followed a validated algorithm using particular ICD-10-CA codes, to identify (and index) adult SCI patients from Ambulatory and Inpatient records between April 1, 2006 and March 31, 2016.&#x0D; Indexed patients were linked to various databases (inpatient, ambulatory, physician claims, provincial insurance registry), and analyzed in multiple perspectives such as demographics patterns, deaths, resource and cost utilization, geographic distribution, and care equity between groups.&#x0D; ResultsThrough 10 years of data we have identified 5217 SCI patients (3309 TSCI; 1908 NTSCI). 68.7% TSCI and 58.6% NTSCI are male. NTSCI patients are approximately 10 years older (46.3 TSCI; 54.5 NTSCI), and have a 3-point higher Charlson score. 1-year mortality in NTSCI is approximately 2.4 times the TSCI group.&#x0D; Hospitalizations, ER visits, critical care time have also been examined. Patients with NTSCI had a higher median index LOS (14 days IQR (4-51)) compared to the traumatic group who had much higher variability (11 days IQR (11-65.5)). Noted 13.7% NTSCI patients and 19.5% TSCI do not have hospitalizations after index (a diverse characteristic of SCI). Resource Intensity Weights, physician billing, rural-urban area utilization have also been compared between the sub-groups.&#x0D; Conclusion/ImplicationsWith the use of administrative databases and a validated algorithm, we described a diverse patient cohort with two main sub-groups (TSCI/NTSCI). Both groups were analyzed upon multiple topics and showed variations. Our results have provided updated knowledge of a comprehensive SCI population in Alberta, Canada, and may lead to improvements on care-giving model.

Identifying and Prioritizing Low Value Care in British Columbia Using Three Administrative Health Data Assets

IntroductionClinical recommendations and/or lists of low value care (i.e., health technologies that provide little to clinical benefit for certain patient groups) have garnered attention internationally through campaigns such as Choosing Wisely. However, uptake of such recommendations at the healthcare system-level remains challenging in the absence of routine, data-driven processes.&#x0D; Objectives and ApproachThe objective of this work was to develop and implement a process, leveraging administrative health data assets and lists of ‘low value’ care, to identify and prioritize technologies at the healthcare system-level for reassessment and potential disinvestment. The British Columbia (BC) healthcare system was selected as the pilot site to test the process. Three provincial administrative health databases were used to examine the extent of low value care across the system: the discharge abstract database (DAD); the Medical Service Plan (MSP) physician claims database; and the MSP laboratory database.&#x0D; ResultsOver 1300 recommendations of low value technologies (i.e., from the National Institute for Health and Care Excellence “do not do” recommendations, low value technologies in the Australian Medical Benefits Schedule, and Choosing Wisely “Top 5” lists) were identified. Using appropriate coding systems for BC’s administrative health data (e.g., International Classification of Diseases), low value technologies were queried to examine frequencies and costs of technology use between fiscal years 2010/11 and 2014/15. This information was used to rank technologies based high budgetary impact, defined as total in-hospital and claims expenditures exceeding $1M in any fiscal year examined. Clinical experts reviewed the ranked technologies prior to dissemination and stakeholder action. Pilot testing resulted in the prioritization of 9 candidate technologies for reassessment in the BC healthcare system.&#x0D; Conclusion/ImplicationsThis work demonstrates the feasibility and strength of using administrative data to identify low value care at the healthcare system-level and prioritize candidates for reassessment. Faced with increasing pressure to control exorbitant costs, while maintaining quality of care, this process has been adopted and operationalized by the BC Ministry of Health.

The Use of Long-Acting Injectable Antipsychotic Therapy for Schizophrenia

IntroductionAntipsychotic medications form the cornerstone of schizophrenia treatment. However, only a minority of patients adhere to their initial antipsychotic regimen. It’s expected that Long-Acting Injectable (LAI) antipsychotics improves patient adherence to treatment, however previous research comparing the use of first generation LAI’s against oral antipsychotics reported results that were inconclusive.&#x0D; Objectives and ApproachExplore the effectiveness of the use of LAI’s in the delivery of mental health services in Alberta.&#x0D; Using linked data from AHS Analytics:&#x0D; &#x0D; Physician claims&#x0D; National Ambulatory Care Reporting System (NACRS),&#x0D; Discharge Abstract database (DAD)&#x0D; Pharmacy Information Network (PIN)&#x0D; Alberta Provincial Registry data&#x0D; &#x0D; &#x0D; Define a cohort of patients on antipsychotic medications.&#x0D; Explore and contrast outcomes related to the use of LAIs against other antipsychotic medication types. Specifically using linked data to define:&#x0D; &#x0D; Treatment Adherence&#x0D; Utilization of LAI vs. other medication&#x0D; Demographic differences&#x0D; Outcomes pre- and post-LAI treatment&#x0D; &#x0D; &#x0D; ResultsA patient cohort was established containing only cases from April 1, 2013 to March 31, 2015. Additional data was used to perform a two year washout and a one year follow-up.&#x0D; Case and medication definitions were determined by a team of psychiatric clinicians. Patient comorbidity information was extracted using previously validated methods.&#x0D; Overall, 6349 incident cases were identified.&#x0D; Preliminary analysis indicate:&#x0D; &#x0D; Overall patient cohort is older than expected&#x0D; Use of additional medication types is correlated with greater health services utilization after diagnosis&#x0D; Patients on only oral medications appear to have lower treatment adherence&#x0D; Males seem to have higher treatment adherence than females&#x0D; No significant differences were found between patients with rural vs. urban postal codes&#x0D; &#x0D; Conclusion/ImplicationsWe faced significant challenges when defining cases, medication use and outcomes. However, the linkage of a large number of data sources gives us powerful and multi-faceted insight into the use of antipsychotic medication use in Alberta. Future work will include work on definition validations and deeper analysis of outcomes.