Validation of case-ascertainment algorithms using health administrative data to identify people who inject drugs in Ontario, Canada

OBJECTIVE

Health administrative data can be used to improve the health of people who inject drugs by informing public health surveillance and program planning, monitoring, and evaluation. However, methodological gaps in the use of these data persist due to challenges in accurately identifying injection drug use at the population level. In this study, we validated case-ascertainment algorithms for identifying people who inject drugs using health administrative data in Ontario, Canada.

STUDY DESIGN AND SETTING

Data from cohorts of people with recent (past 12 month) injection drug use, including those participating in community-based research studies or seeking drug treatment were linked to health administrative data in Ontario from 1992-2020. We assessed the validity of algorithms to identify injection drug use over varying lookback periods (i.e., all years of data [1992 onwards] or within the past 1-5 years), including inpatient and outpatient physician billing claims for drug use, emergency department visits or hospitalizations for drug use or injection-related infections, and opioid agonist treatment (OAT).

RESULTS

Algorithms were validated using data from 15,241 people with recent IDU (918 in community cohorts, 14,323 seeking drug treatment). An algorithm consisting of ≥1 physician visit, emergency department visit or hospitalization for drug use, or OAT record could effectively identify IDU history (91.6% sensitivity, 94.2% specificity) and recent IDU (using 3 years lookback: 80.4% sensitivity, 99% specificity) among community cohorts. Algorithms were generally more sensitive among people who inject drugs seeking drug treatment.

CONCLUSION

Validated algorithms using health administrative data performed well in identifying people who inject drugs. Despite high sensitivity and specificity, the positive predictive value of these algorithms will vary depending on the underlying prevalence of injection drug use in the population in which they are applied.

Eleven-year trends in socioeconomic inequalities in the prevalence and incidence of pharmacologically treated type 2 diabetes in France, 2010-2020

AIM

This study aimed to describe the association between socioeconomic inequalities and the prevalence and incidence of pharmacologically-treated type 2 diabetes in European France over the 2010-2020 period.

METHODS

Diabetes cases were identified using a validated algorithm from the French National Health Data System. Analysis was restricted to adults aged 45 years and older to focus on type 2 diabetes. Socioeconomic inequalities were measured for all years in European France using the French deprivation index (FDep, 2015 version), which is an area-based deprivation indicator using population-weighted quintiles (Q1 corresponds to the least deprived municipalities). The relative risks of diabetes prevalence and incidence associated with FDep quintiles (Q1 as the reference) were estimated by sex using a log-linear Poisson model adjusted for year, age and French department. The study population was the French health consumers aged 45 years and over (from 24,228,526 in 2010 to 29,772,928 in 2020).

RESULTS

A positive gradient was observed in the relative risks of type 2 diabetes prevalence and incidence by FDep quintiles over the study period. The strength of the estimated associations increased over the last decade for prevalence among men and women and for incidence among men in the two most deprived quintiles.

CONCLUSION

Thus, type 2 diabetes prevention should include a proportionate universalism approach, proposing actions of greater intensity in the most deprived areas.

Disparities in alcohol- and substance-related hospitalizations and deaths across sexual orientations in Canada: a longitudinal study

OBJECTIVES

We aim to quantify disparities in substance-related acute events (i.e. hospitalizations and deaths for substances, including alcohol, cannabis, opioids, narcotics, and/or illicit drugs) across sexual orientations based on health administrative data.

STUDY DESIGN

This was a longitudinal analysis from six waves of the Canadian Community Health Survey (2009-2014) linked to hospitalization/mortality data in Ontario, Canada.

METHODS

The study sample consisted of survey respondents aged ≥15 (weighted n = 15,406,000) who provided a valid response to the question about sexual orientation. The primary outcome was substance-related acute events, including hospitalizations and deaths due to alcohol, cannabis, opioids, narcotics, and/or illicit drugs. Disparities in substance-related acute events across sexual orientation and gender were examined using flexible parametric survival analysis.

RESULTS

Bisexual women had hazard ratios of 2.46 (95% confidence interval: 1.46-4.15) for any substance-related acute event and 2.67 (95% confidence interval: 1.42-5.00) for non-alcohol substance acute events compared to heterosexual women. Lesbian women did not exhibit significant differences in acute event risk compared to heterosexual women. Gay and bisexual men demonstrated elevated but not statistically significant risks compared to heterosexual men.

CONCLUSIONS

Bisexual women face higher risks of substance-related acute events, potentially due to self-medication of unique stressors brought on by discrimination and isolation. Enhanced education and training for healthcare professionals are essential to increase awareness and sensitivity toward the heightened substance use risk among lesbian, gay, and bisexual individuals. Targeted interventions aimed at reducing substance use problems among bisexual individuals warrant increased funding and research.

Validation of ICD-10 codes for studying foreign body airway obstructions: A health administrative data cohort study

Aim

To validate a case definition for foreign body airway obstructions (FBAO) using International Classification of Diseases version 10 (ICD-10) codes to accurately identify patients in administrative health databases and improve reporting on this injury.

Methods

We identified prehospital patient encounters in Alberta, Canada between Jan 1, 2018 and Dec 31, 2021 by querying the provincial emergency medical services' (EMS) patient care records for FBAO-related presentations, EMS protocols, or treatments. We deterministically linked EMS patient encounters to data on emergency department visits and hospital admissions, which included ICD-10 codes. Two physicians independently reviewed encounters to determine true FBAO cases. We then calculated diagnostic accuracy measures (sensitivity, specificity, likelihood ratios) of various algorithms.

Results

We identified 3677 EMS patient encounters, 2121 were linked to hospital administrative databases. Of these encounters, 825 (38.9%) were true FBAO. The combination of two ICD-10 codes (T17 = foreign body in the respiratory tract or T18.0 = foreign body in the mouth) was the most specific algorithm (96.9% [95%CI 95.8-97.8%]), while the combination of all FBAO-related ICD-10 codes and R06.8 (other breathing abnormalities) was the most sensitive (75.0% [95%CI 71.9-78.0]). We identified an additional 453 (35.4%) FBAO cases not transported by EMS (due to death or transport refusal), and therefore not linked to the hospital administrative databases. Of these unlinked encounters, 44 (9.7%) cases resulted in the patient's death.

Conclusions

FBAO can be identified with reasonable accuracy using health administrative data and ICD-10 codes. All algorithms had a trade-off between sensitivity and specificity, and failed to identify a third of FBAO cases, of which 10% resulted in death.

Evaluation of audit and feedback to family physicians on prescribing of opioid analgesics to opioid-naïve patients: A pragmatic randomized delay trial

BACKGROUND

Exposure to opioid analgesics have historically raised concern for a risk of developing opioid use disorder. Prescriber audit-and-feedback interventions may reduce opioid prescribing, but some studies have shown detrimental effects for current users. We examined the effectiveness of an audit and feedback intervention, named Portrait, to reduce initiation of opioid analgesics among opioid-naïve patients experiencing pain.

METHODS

REDONNA was a single-blinded, two-arm (Early vs Delayed mailing) randomized trial of a portrait for eligible family physicians (FPs) in British Columbia (BC), Canada. The primary outcome was the change in the number of initiations of opioid analgesic prescriptions written by FPs for acute/chronic pain management. We compared outcomes for a 6-month window before vs. after each mailed intervention, using differences in percent differences (DPD) with 95% confidence intervals (CI) and odds ratios (OR) from logistic regressions adjusted for clustering of patients by FP.

RESULTS

In the Early (n = 2260) and Delayed (n = 2156) groups, opioid initiations per month were the same in the Before (2.10 Early; 2.06 Delayed) and After (1.94 Early; 1.95 Delayed) windows. The DPD was -2.1% (CI: -4.4% to 0.3%), and ORs were: 0.98 (CI: 0.96 to 1.01) for any opioid, 0.97 (CI: 0.94 to 1.01) for codeine (62% of initiations), and 1.0 (CI: 0.97 to 1.07) for tramadol (25% of initiations). There were no differences in mean quantity of tablets, mean milligrams of morphine equivalents (MME), or mean number of days.

CONCLUSION

Portrait had no impact on FPs' rates of prescribing opioid analgesics to opioid-naïve patients experiencing pain.

TRIAL REGISTRATION

The study was registered prospectively on 30 March 2020 at the ISRCTN Register (https://www.isrctn.com/ISRCTN34246811).

The burden of psychiatric disorders associated with orofacial cleft pathology among children in Ontario, Canada

BACKGROUND

Individuals with orofacial cleft (OFC) may be at a higher risk of developing psychiatric disorders (PD) than the general population. We determined the risk of psychiatric diagnoses in children with OFC in Canada.

METHODS

This population-based retrospective cohort study used health administrative data from the province of Ontario, Canada. Children with OFC who were born between April 1, 1994, and March 31, 2017, in Ontario were matched to five non-OFC children based on sex, date of birth, and mother's age. We determined the rate of events and time-to-event for first diagnosis of PD in children aged ≥ 3 years (y), and for intellectual developmental delay (IDD) from birth. Risk factors for PD and IDD were assessed using 1-way ANOVA for means, Kruskal-Wallis for medians, and the χ2 test for categorical variables.

OUTCOMES

There were 3051 children with OFC (matched to 15,255 controls), of whom 2515 patients with OFC (12,575 controls) had a complete follow-up to the third birthday. Children with OFC were more likely to have PD than controls (54.90 vs. 43.28 per 1000 patient-years, P < .001), with a mean age to first diagnosis of 8.6 ± 4.2 y. The cleft palate group had the highest risk (HR 1.33, 95% CI 1.18-1.49). Children with OFC also had a higher risk of IDD than non-OFC children (27.78 vs. 3.46 per 1000 patient-years, p < .001).

INTERPRETATION

Children born with OFC in Ontario had a higher risk of psychiatric diagnosis and IDD compared to controls. Further research is also required to better understand the predictors of variation in risk, including geographic location and the presence of congenital abnormalities, and identify potential areas for intervention.

EVIDENCE RATING SCALE FOR PROGNOSTIC/RISK STUDIES

Level II.

Differences in clinical presentation at first hospitalization and the impact on involuntary admissions among first-generation migrant groups with non-affective psychotic disorders

BACKGROUND

Some migrant and ethnic minority groups have a higher risk of coercive pathways to care; however, it is unclear whether differences in clinical presentation contribute to this risk. We sought to assess: (i) whether there were differences in clinician-rated symptoms and behaviours across first-generation immigrant and refugee groups at the first psychiatric hospitalization after psychosis diagnosis, and (ii) whether these differences accounted for disparities in involuntary admission.

METHODS

Using population-based health administrative data from Ontario, Canada, we constructed a sample (2009-2013) of incident cases of non-affective psychotic disorder followed for two years to identify first psychiatric hospitalization. We compared clinician-rated symptoms and behaviours at admission between first-generation immigrants and refugees and the general population, and adjusted for these variables to ascertain whether the elevated prevalence of involuntary admission persisted.

RESULTS

Immigrants and refugee groups tended to have lower ratings for affective symptoms, self-harm behaviours, and substance use, as well as higher levels of medication nonadherence and poor insight. Immigrant groups were more likely to be perceived as aggressive and a risk of harm to others, and both groups were perceived as having self-care issues. Adjustment for perceived differences in clinical presentation at admission did not attenuate the higher prevalence of involuntary admission for immigrant and refugee groups.

CONCLUSIONS

First-generation migrant groups may differ in clinical presentation during the early course of psychotic illness, although these perceived differences did not explain the elevated rates of involuntary admission. Further research using outpatient samples and tools with established cross-cultural validity are warranted.

Validation of Administrative Health Data Algorithms for Identifying Persons with Parkinson's Disease and the 10-Year Prevalence Trend in Bologna, Italy

INTRODUCTION

Health administrative databases are widely used for the estimation of the prevalence of Parkinson's disease (PD). Few in general, and none used in Italy, have been validated by testing their diagnostic accuracy. The primary objective was to validate two algorithms for the identification of persons with PD using clinical diagnosis as the reference standard on an Italian sample of people with PD. The second objective was to estimate 10-year trends in PD prevalence in the Bologna Local Health Trust from 2010 to 2019.

METHODS

Two algorithms (index tests) applied to health administrative databases (hospital discharge, drug prescriptions, exemptions for medical costs) were validated against clinical diagnosis of PD by an expert neurologist (reference standard) in a cohort of consecutive outpatients. Sensitivity and specificity with relative 95% confidence intervals (CIs) were calculated. The prevalence of PD in a specific year was estimated as the ratio between the number of subjects fulfilling any criteria of the algorithm with better diagnostic accuracy and the total population in the same year (×1,000), stratified by age, sex, and district of residence.

RESULTS

The two algorithms showed high accuracy for identifying patients with PD: one with greater sensitivity of 94.2% (CI: 88.4-97.6) and the other with greater specificity of 98.1% (CI: 97.7-98.5). For the estimation of prevalence, we chose the most specific algorithm with the fewest total number of misclassified cases. We identified 3,798 people with PD as of December 31, 2019, corresponding to a prevalence of 4.3 per 1,000 inhabitants (CI: 4.2-4.4). Prevalence was higher in males (4.7, CI: 4.5-5.0) than females (3.8, CI: 3.7-4.0) and increased with age. The crude prevalence over time was slightly elevated as it followed a progressive aging of the population. When stratifying the prevalence for age groups, we did not observe a trend except in the 45-64 year category where we observed an increasing trend over time.

CONCLUSION

Algorithms based on administrative data are accurate when detecting people with PD in the Italian public health system. In a large northern Italian population, increased prevalence of about 10% was observed in the decade 2010-2019 and is explained by increased life expectancy. These data may be useful in planning the allocation of health care resources for people with PD.

Systemic autoimmune rheumatic diseases and multiple industrial air pollutant emissions: A large general population Canadian cohort analysis

BACKGROUND

Past investigations of air pollution and systemic autoimmune rheumatic diseases (SARDs) typically focused on individual (not mixed) and overall environmental emissions. We assessed mixtures of industrial emissions of fine particulate matter (PM_2.5), nitrogen dioxide (NO₂), and sulfur dioxide (SO₂) and SARDs onset in Ontario, Canada.

METHODS

We assembled an open cohort of over 12 million adults (without SARD diagnoses at cohort entry) based on provincial health data for 2007-2020 and followed them until SARD onset, death, emigration, or end of study (December 2020). SARDs were identified using physician billing and hospitalization diagnostic codes for systemic lupus, scleroderma, myositis, undifferentiated connective tissue disease, and Sjogren's. Rheumatoid arthritis and vasculitis were not included. Average PM_2.5, NO₂, and SO₂ industrial emissions from 2002 to one year before SARDs onset or end of study were assigned using residential postal codes. A quantile g-computation model for time to SARD onset was developed for the industrial emission mixture, adjusting for sex, age, income, rurality index, chronic obstructive pulmonary disease (as a proxy for smoking), background (environmental overall) PM_2.5, and calendar year. We conducted stratified analyses across age, sex, and rurality.

RESULTS

We identified 43,931 new SARD diagnoses across 143,799,564 person-years. The adjusted hazard ratio for SARD onset for an increase in all emissions by one decile was 1.018 (95% confidence interval 1.013-1.022). Similar positive associations between SARDs and the mixed emissions were observed in most stratified analyses. Industrial PM_2.5 contributed most to SARD risk.

CONCLUSIONS

Industrial air pollution emissions were associated with SARDs risk.

Patient and Physician Factors Associated with First Diagnosis of Non-affective Psychotic Disorder in Primary Care

Primary care physicians play a central role in pathways to care for first-episode psychosis, and their increased involvement in early detection could improve service-related outcomes. The aim of this study was to estimate the proportion of psychosis first diagnosed in primary care, and identify associated patient and physician factors. We used linked health administrative data to construct a retrospective cohort of people aged 14-35 years with a first diagnosis of non-affective psychosis in Ontario, Canada between 2005-2015. We restricted the sample to patients with help-seeking contacts for mental health reasons in primary care in the six months prior to first diagnosis of psychotic disorder. We used modified Poisson regression models to examine patient and physician factors associated with a first diagnosis of psychosis in primary care. Among people with early psychosis (n = 39,449), 63% had help-seeking contacts in primary care within six months prior to first diagnosis. Of those patients, 47% were diagnosed in primary care and 53% in secondary/tertiary care. Patients factors associated with lower likelihood of diagnosis in primary care included male sex, younger age, immigrant status, and comorbid psychosocial conditions. Physician factors associated with lower likelihood of diagnosis in primary care included solo practice model, urban practice setting, international medical education, and longer time since graduation. Our findings indicate that primary care is an important contact for help-seeking and diagnosis for a large proportion of people with early psychosis. For physicians less likely to diagnose psychosis in primary care, targeted resources and interventions could be provided to support them in caring for patients with early psychosis.

Real world evidence of the association between medication and life expectancy in elderly inflammatory bowel disease: a population-based cohort study

BACKGROUND

Life expectancy in people with inflammatory bowel disease (IBD) has increased but remains shorter than in people without IBD. We describe the life expectancy associated with IBD therapies among the growing number of older adults living with IBD.

METHODS

Older adults (≥ 65 years) with IBD were identified from population-based health administrative data using a validated algorithm. Life expectancy on patients' 65th birthday, stratified by sex, was calculated using a period life table approach from age- and sex-specific mortality rates among patients receiving immunomodulator monotherapy, biologic monotherapy, combination therapy, mesalamine, systemic steroids, and no therapy.

RESULTS

Among 28,260 older adults with IBD (239,125 person-years of follow-up), life expectancy at 65 years was longest for patients taking mesalamine (females: 22.1 years, 95% CI 21.8-22.5; males: 19.6 years, 95% CI 19.3-20.0) and shortest for patients taking steroids (females: 11.7 years, 95% CI 11.0-12.4; males 10.3 years, 95% CI 9.7-10.8). Life expectancy was similar for patients receiving immunomodulator monotherapy and biologic monotherapy. Immunomodulator monotherapy was associated with a reduction in life expectancy compared to combination therapy by 5.1 (95% CI 2.3-7.8) in females and 2.8 years (95% CI 0.1-5.5) in males.

CONCLUSIONS

Life expectancy varies across therapies used for IBD, with differences likely arising from a combination of medication effectiveness, safety profiles, disease severity, and comorbid conditions. These considerations should be balanced when deciding on a therapeutic approach for the management of IBD in older adults.

Validating viral hepatitis B and C diagnosis codes: a retrospective analysis using Ontario's health administrative data

OBJECTIVE

We aimed to determine the criterion validity of using diagnosis codes for hepatitis B virus (HBV) and hepatitis C virus (HCV) to identify infections.

METHODS

Using linked laboratory and administrative data in Ontario, Canada, from January 2004 to December 2014, we validated HBV/HCV diagnosis codes against laboratory-confirmed infections. Performance measures (sensitivity, specificity, and positive predictive value) were estimated via cross-validated logistic regression and we explored variations by varying time windows from 1 to 5 years before (i.e., prognostic prediction) and after (i.e., diagnostic prediction) the date of laboratory confirmation. Subgroup analyses were performed among immigrants, males, baby boomers, and females to examine the robustness of these measures.

RESULTS

A total of 1,599,023 individuals were tested for HBV and 840,924 for HCV, with a resulting 41,714 (2.7%) and 58,563 (7.0%) infections identified, respectively. HBV/HCV diagnosis codes ± 3 years of laboratory confirmation showed high specificity (99.9% HBV; 99.8% HCV), moderate positive predictive value (70.3% HBV; 85.8% HCV), and low sensitivity (12.8% HBV; 30.8% HCV). Varying the time window resulted in limited changes to performance measures. Diagnostic models consistently outperformed prognostic models. No major differences were observed among subgroups.

CONCLUSION

HBV/HCV codes should not be the only source used for monitoring the population burden of these infections, due to low sensitivity and moderate positive predictive values. These results underscore the importance of ongoing laboratory and reportable disease surveillance systems for monitoring viral hepatitis in Ontario.

Individualized prescribing portraits to reduce inappropriate initiation of opioid analgesics to opioid naïve patients in primary care: Protocol for a randomized controlled trial

Background Opioid analgesics are frequently initiated for chronic and acute pain despite weak evidence of benefit, although prescribing rates of some analgesics decreased in the context of the epidemic. In some populations, up to a quarter of opioid naïve persons prescribed opioids for non-cancer pain develop prescription opioid use disorder (OUD). Audit and feedback interventions rely on constructive use of routinely collected data to align professional behaviours and clinical practice with best evidence. These interventions have been shown to help reduce inappropriate initiation. However, effectiveness and acceptability of individualized "portraits" of physicians' prescribing patterns, to reduce inappropriate initiation of opioid analgesics to opioid naïve persons, have not been evaluated. Methods REDONNA is a mixed-methods randomized study testing the effectiveness of individualized prescribing Portraits to reduce inappropriate initiation of opioid analgesics. This intervention to improve safety of opioid prescribing in primary care in British Columbia (BC), Canada involves mailing individual prescribing portraits to an 'early group' of 2604 family physicians, followed in 6 months by a mailing to 2553 family physicians in the 'delayed group'. Primary outcome is number of new opioid prescriptions initiated in opioid naïve people, measured using administrative data from a centralized medication monitoring database covering all prescription opioids dispensed from BC community pharmacies. Secondary endpoints will compare prescribing impact between the two groups. A qualitative sub-study will examine feasibility among a purposive sample of physicians and patients. Discussion This trial provides important evidence on the intervention's potential to steer policy and practice on inappropriate opioid analgesics initiation. Trial registration: The study was registered prospectively on 30 March 2020 at the ISRCTN Register (https://www.isrctn.com/ISRCTN34246811).

Validation of an Algorithm to Detect Multiple Sclerosis Cases in Administrative Health Databases in Piedmont (Italy): An Application to the Estimate of Prevalence by Age and Urbanization Level

INTRODUCTION

Italy is considered a high-risk country for multiple sclerosis (MS). Exploiting electronic health archives (EHAs) is highly useful to continuously monitoring the prevalence of the disease, as well as the care delivered to patients and its outcomes. The aim of this study was to validate an EHA-based algorithm to identify MS patients, suitable for epidemiological purposes, and to estimate MS prevalence in Piedmont (North Italy).

METHODS

MS cases were identified, in the period between January 1, 2012 and December 31, 2017, linking data from 4 different sources: hospital discharges, drug prescriptions, exemptions from co-payment to health care, and long-term care facilities. Sensitivity of the algorithm was tested through record linkage with a cohort of 656 neurologist-confirmed MS cases; specificity was tested with a cohort of 2,966,293 residents presumably not affected by MS. Undercount was estimated by a capture-recapture method. We calculated crude, and age- and gender-specific prevalence. We also calculated age-adjusted prevalence by level of urbanization of the municipality of residence.

RESULTS

On December 31, 2017, the algorithm identified 8,850 MS cases. Sensitivity was 95.9%, specificity was 99.97%, and the estimated completeness of ascertainment was 91.9%. The overall prevalence, adjusted for undercount, was 152 per 100,000 among men and 286 among women; it increased with increasing age and reached its peak value in the 45- to 54-year class, followed by a progressive reduction. The age-adjusted prevalence of residents in cities was 15% higher than in those living in the countryside.

DISCUSSION/CONCLUSION

We validated an algorithm based on EHAs to identify cases of MS for epidemiological use. The prevalence of MS, adjusted for undercount, was among the highest in Italy. We also found that the prevalence was higher in highly urbanized areas.

Factors Associated with Stroke Coding Quality: A Comparison of Registry and Administrative Data

BACKGROUND

The International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM) codes are commonly used to identify patients with diseases or clinical conditions for epidemiological research. We aimed to determine the diagnostic agreement and factors associated with a clinician-assigned stroke diagnosis in a national registry and the ICD-10-AM codes recorded in government-held administrative data.

MATERIALS AND METHODS

Data from 39 hospitals (2009-2013) participating in the Australian Stroke Clinical Registry (AuSCR) were linked and merged with person-level administrative data. The AuSCR clinician-assigned stroke diagnosis was the reference standard. Concordance was defined as agreement between the clinician-assigned diagnosis and the ICD-10-AM codes for acute stroke or transient ischemic attack (TIA) (ICD-10-AM codes: I61-I64, G45.9). Multivariable logistic regression was undertaken to assess factors associated with coded diagnostic concordance.

RESULTS

A total of 14,716 patient admissions were included (46% female, 63% ischemic, 14% intracerebral hemorrhage [ICH], 18% TIA and 5% unspecified stroke based on the reference standard). Principal ICD-10-AM code concordance was ICH: 76.7%; ischemic stroke: 72.2%; TIA: 80.2%; unspecified stroke: 50.8%. Factors associated with a greater odds of ischemic stroke concordance included: treatment in a stroke unit (adjusted Odds Ratio, aOR:1.58; 95% confidence interval (CI) 1.37, 1.82); length of stay >4 days (aOR:1.30; 95% CI 1.17, 1.45); and discharge destination other than home (Residential care aOR:1.57; 95% CI 1.24, 1.96; Inpatient rehabilitation aOR:1.63; 95% CI 1.43, 1.86).

CONCLUSIONS

Diagnostic concordance varied based on stroke type. Future research to improve the quality of coding for stroke should focus on patients not treated in stroke units or with shorter lengths of stay where documentation in medical records may be limited.

Comparing health insurance data and health interview survey data for ascertaining chronic disease prevalence in Belgium

Background

Health administrative data were increasingly used for chronic diseases (CDs) surveillance purposes. This cross sectional study explored the agreement between Belgian compulsory health insurance (BCHI) data and Belgian health interview survey (BHIS) data for asserting CDs.

Methods

Individual BHIS 2013 data were linked with BCHI data using the unique national register number. The study population included all participants of the BHIS 2013 aged 15 years and older. Linkage was possible for 93% of BHIS-participants, resulting in a study sample of 8474 individuals. For seven CDs disease status was available both through self-reported information from the BHIS and algorithms based on ATC-codes of disease-specific medication, developed on demand of the National Institute for Health and Disability Insurance (NIHDI). CD prevalence rates from both data sources were compared. Agreement was measured using sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) assuming BHIS data as gold standard. Kappa statistic was also calculated. Participants’ sociodemographic and health status characteristics associated with agreement were tested using logistic regression for each CD.

Results

Prevalence from BCHI data was significantly higher for CVDs but significantly lower for COPD and asthma. No significant difference was found between the two data sources for the remaining CDs. Sensitivity was 83% for CVDs, 78% for diabetes and ranged from 27 to 67% for the other CDs. Specificity was excellent for all CDs (above 98%) except for CVDs. The highest PPV was found for Parkinson’s disease (83%) and ranged from 41 to 75% for the remaining CDs. Irrespective of the CDs, the NPV was excellent. Kappa statistic was good for diabetes, CVDs, Parkinson’s disease and thyroid disorders, moderate for epilepsy and fair for COPD and asthma. Agreement between BHIS and BCHI data is affected by individual sociodemographic characteristics and health status, although these effects varied across CDs.

Conclusions

NHIDI’s CDs case definitions are an acceptable alternative to identify cases of diabetes, CVDs, Parkinson’s disease and thyroid disorders but yield in a significant underestimated number of patients suffering from asthma and COPD. Further research is needed to refine the definitions of CDs from administrative data.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13690-020-00500-4.

Engagement in primary health care among marginalized people who use drugs in Ottawa, Canada

BACKGROUND

There may be less primary health care engagement among people who use drugs (PWUD) than among the general population, even though the former have greater comorbidity and more frequent use of emergency department care. We investigated factors associated with primary care engagement among PWUD.

METHODS

The Participatory Research in Ottawa: Understanding Drugs (PROUD) cohort study meaningfully engaged and trained people with lived experience to recruit and survey marginalized PWUD between March-December 2013. We linked this survey data to provincial-level administrative databases held at ICES. We categorized engagement in primary care over the 2 years prior to survey completion as: not engaged (< 3 outpatient visits to the same family physician) versus engaged in care (3+ visits to the same family physician). We used multivariable logistic regression to determine factors associated with engagement in primary care.

RESULTS

Characteristics of 663 participants included a median age of 43 years, 76% men, and 67% living in the two lowest income quintile neighborhoods. Despite high comorbidity and a median of 4 (interquartile range 0-10) primary care visits in the year prior to survey completion, only 372 (56.1%) were engaged in primary care. Engagement was most strongly associated with the following factors: receiving provincial benefits, including disability payments (adjusted odds ratio [AOR] 4.14 (95% confidence interval [CI] 2.30 to 7.43)) or income assistance (AOR 3.69 (95% CI 2.00 to 6.81)), having ever taken methadone (AOR 3.82 (95% CI 2.28 to 6.41)), mental health comorbidity (AOR 3.43 (95% CI 2.19 to 5.38)), and having stable housing (AOR 2.09 (95% CI 1.29 to 3.38)).

CONCLUSIONS

Despite high comorbidity, engagement in primary care among PWUD was low. Our findings suggest that social care (housing, disability, and income support) and mental health care are associated with improved primary care continuity; integration of these care systems with primary care and opioid substitution therapy may lessen the significant morbidity and acute care use among PWUD.

Long-term Outcomes of Patients Surgically Treated for Hirschsprung Disease

Background

The only curative treatment for Hirschsprung disease (HD) is surgical repair. However, some patients experience poor postoperative outcomes. We determined long-term outcomes of all HD patients in Ontario, Canada’s most populous province.

Methods

We conducted a retrospective cohort study including all children with HD born between April 1, 1991 and March 31, 2014 in Ontario using linked health administrative data. Each HD case was matched to five non-HD controls on sex, date of birth, region of residence and income and followed to March 31, 2016. Chronic diarrhea and constipation were identified using combinations of outpatient physician billing codes in both HD patients and non-HD residents of the province. We determined risk factors associated with diarrhea and constipation, including surgery type and sociodemographic characteristics, using multivariable conditional logistic regression, and reported adjusted odds ratios (aORs).

Results

There were 3,265,172 children born in the study period, of whom 673 had HD. Compared to controls, chronic constipation was more common in HD patients (27.5% versus 2.1%; aOR 17.2, 95% CI 12.6 to 23.4), as was chronic diarrhea (29.9% versus 6.9%, aOR 5.22, 95% CI 4.19 to 6.50). In HD patients, older age at surgery was associated with increased risk of chronic constipation (OR 2.71, 95% CI 1.75 to 4.20). Surgery type, sex, rural/urban residence and income were not associated with risk of chronic constipation or diarrhea.

Conclusion

Chronic constipation and diarrhea were common following surgery for HD. Older age at surgery was associated with subsequent risk of chronic constipation. Surgery type was not associated with increased risk of chronic constipation or diarrhea.

External validation demonstrated the Ottawa SAH prediction models can identify pSAH using health administrative data

OBJECTIVES

The objective of the study is to externally validate three primary subarachnoid hemorrhage (pSAH) identification models.

STUDY DESIGN AND SETTING

We evaluated three models that identify pSAH using recursive partitioning (A), logistic regression (B), and a prevalence-adjusted logistic regression(C), respectively. Blinded chart review and/or linkage to existing registries determined pSAH status. We included all patients aged ≥18 in four participating center registries or whose discharge abstracts contained ≥1 administrative codes of interest between January 1, 2012 and December 31, 2013.

RESULTS

A total of 3,262 of 193,190 admissions underwent chart review (n = 2,493) or registry linkage (n = 769). A total of 657 had pSAH confirmed (20·1% sample, 0·34% admissions). The sensitivity, specificity, and positive predictive value (PPV) were as follows: i) model A: 98·3% (97·0-99·2), 53·5% (51·5-55·4), and 34·8% (32·6-37·0); ii) model B (score ≥6): 98·0% (96·6-98·9), 47·4% (45·5-49·4), and 32·0% (30·0-34·1); and iii) model C (score ≥2): 95·7% (93·9-97·2), 85·5% (84·0-86·8), and 62·3 (59·3-65·3), respectively. Model C scores of 0, 1, 2, 3, or 4 had probabilities of 0·5% (0·2-1·5), 1·5% (1·0-2·2), 24·8% (21·0-29·0), 90·0% (86·8-92·0), and 97·8% (88·7-99·6), without significant difference between centers (P = 0·86). The PPV of the International Classification of Diseases code (I60) was 63·0% (95% confidence interval: 60·0-66·0).

CONCLUSIONS

All three models were highly sensitive for pSAH. Model C could be used to adjust for misclassification bias.

Identifying incident Parkinson's disease using administrative diagnostic codes: a validation study

BACKGROUND

Administrative databases that capture diagnostic codes are increasingly being used worldwide for research because they can save time and reduce costs. However, assessing validity is necessary before defining diseases using only diagnostic codes in research applications.

OBJECTIVE

Our objective was to assess the validity of using diagnostic codes to identify incident Parkinson's disease (PD) cases in Olmsted County, Minnesota using an established standard for comparison (1976-2005).

METHODS

Cases were identified solely using computer programs applied to administrative diagnostic code indexes from the Rochester Epidemiology Project (REP). Two codes >30 days apart or one code on the death certificate constituted PD. The standard was a clinical diagnosis by movement disorders specialists based on medical record review. Validity was assessed using positive predictive value (PPV) and sensitivity. Numbers of incident cases and incidence rates were compared between the two ascertainment methods by sex.

RESULTS

The codes only method over-counted the number of incident PD cases by 73% (804 versus 464), and this over-counting generally increased with calendar year. Sensitivity was 80% (95% CI [76%, 84%]) and PPV was 46% (95% CI [34%, 50%]). Disease status misclassification accounted for two-thirds of falsely identified cases, where individuals were found to not have PD (43%) or even parkinsonism (23%) after medical record review. The codes only method also over-estimated the incidence rate time trend for men and women by approximately two-fold.

CONCLUSION

In our context, using administrative diagnostic codes only to identify incident PD cases is not recommended unless more accurate algorithms are developed.

Measuring multimorbidity series. An overlooked complexity - Comparison of self-report vs. administrative data in community-living adults: Paper 3. Agreement across data sources and implications for estimating associations with health service use

OBJECTIVE

The objective of this study is to describe agreement between administrative and self-report data on the number and type of chronic conditions (CCs) and determine whether associations between CC count and health service use differ by data source.

STUDY DESIGN AND SETTING

We linked Canadian Community Health Survey and administrative data for a cohort of adults aged 45+ years in Ontario and identified 12 CCs from both data sources. Agreement was described by count and constituent CCs. We estimated associations between CC count (self-report and administrative data) and health service use (administrative data only) over 1 year.

RESULTS

Among 71,317 adults, 26.9% showed agreement on both count and constituent CCs but agreement declined with increasing CCs. Health service use increased with CC count but the association was stronger when CCs were measured with administrative data. For example, when measured with administrative data, the odds of a general practitioner visit for 5+ CCs vs. none was 20.3 (95% CI 20.0-20.5) but when using self-report data, the estimate was 8.0 (95% CI 7.8-8.2).

CONCLUSION

Agreement on the number of CCs was low and resulted in different estimates on the association with health service use, illustrating the challenges in CC measurement and the ability to interpret the effects on outcomes.

Development of an administrative definition for celiac disease

Objective

The investigation and management of celiac disease places a high burden on the health care system. Accurate methods to ascertain cases of celiac disease (CD) in population-based administrative data can facilitate epidemiologic and health services research to guide disease management. The study aim was to develop and validate administrative data case definitions for CD to facilitate further studies about the effect of CD on osteoporosis and fracture risk.

Results

Population-based data from the Manitoba Bone Mineral Density (BMD) Program registry, which contains medical information on all individuals in the province of Manitoba, Canada who have received BMD testing, was used to define the study cohort. Linked hospital discharge abstracts and physician billing claims were used to ascertain diagnoses of celiac disease in administrative data. A population-based CD serologic registry was used as the validation database. One diagnosis code in hospital discharge abstracts or two or more diagnosis codes in physician billing claims optimized the detection of positive celiac serology with sensitivity of 84% (95% CI 80–88%), specificity of 97% (95% CI 80–88%), PPV of 80% (95% CI 80–88%), and NPV of 97% (95% CI 80–88%). Our administrative data case definition for celiac disease demonstrates good sensitivity and specificity for detecting positive celiac serology.

The probability of diabetes and hypertension by levels of neighborhood walkability and traffic-related air pollution across 15 municipalities in Southern Ontario, Canada: A dataset derived from 2,496,458 community dwelling-adults

Individuals’ risk for cardiovascular disease is shaped by lifestyle factors such as participation in physical activity. Some studies have suggested that rates of physical activity may be higher in walkable neighborhoods that are more supportive of engaging in physical activity in daily life. However, walkable neighborhoods may also contain increased levels of traffic-related air pollution (TRAP). Traffic-related air pollution, often measured through a surrogate marker (e.g. NO₂), has been associated cardiovascular disease risk and risk factors [1], [2], [3], [4]. The higher levels of TRAP in walkable neighborhoods may in turn increase the likelihood of developing conditions like hypertension and diabetes. Our recent work assessed how walkability and TRAP jointly affect the odds of diabetes and hypertension in a sample of community-dwelling adults from Southern Ontario, Canada [5]. This article contains additional data on the probability and odds of hypertension and diabetes according to their walkability and TRAP exposures. Data on cardiovascular risk factors were collected using health administrative databases and environmental exposures were assessed using national land use regression models predicting ground level concentrations of NO₂ and validated walkability indices. The included data were generated using logistic regression accounting for exposures, covariates, and neighborhood clustering. These data may be used as primary data in future health risk assessments and systematic reviews, or to aid in the design of studies examining interactions between built environment and TRAP exposures (e.g. sample size calculations).

Estimating age-specific influenza-associated asthma morbidity in Ontario, Canada

BACKGROUND

There is a need to quantify the potential benefits of influenza-focused interventions in reducing asthma morbidity at a population level. This study aims to estimate age-specific annual excess asthma morbidity attributable to influenza in Ontario, Canada.

METHODS

Weekly counts of hospitalizations, emergency department (ED) visits and outpatient physician office visits for asthma were obtained from health administrative data in Ontario from 2010 to 2015, for ages 0-14, 15-59 and 60+. Asthma morbidity was modelled as a function of influenza A and B activity using linear regression, controlling for seasonal and long-term trend, mean temperature and respiratory syncytial virus. Excess asthma morbidity attributable to influenza was calculated as the difference between full model predictions and model predictions with influenza A and B variables set to 0.

RESULTS

Annually, influenza was associated with the following rates of excess asthma morbidity, per 100,000 people with prevalent asthma: 12.5 hospitalizations for ages 15-59 (95% confidence interval (CI): 1.1-23.5); 35.7 hospitalizations for ages 60+ (95% CI: 3.3-67.1); 114.1 ED visits for ages 15-59 (95% CI: 46.9-181.6); 154.6 ED visits for ages 60+ (95% CI: 86.7-223.3); and 1025.7 outpatient physician office visits for ages 60+ (95% CI: 79.0-1877.3).

CONCLUSIONS

Influenza was associated with excess asthma hospitalizations and ED visits for ages 15-59 and 60+ and outpatient physician office visits for ages 60+. Individuals with asthma aged 15-59 and 60+ might be important targets for influenza-focused interventions, to reduce asthma morbidity at the population level.

Measuring multimorbidity in older adults: comparing different data sources

Background

Multimorbidity is a global health issue, particularly for older adults in the primary care setting. An adequate portrayal of its epidemiology is essential to properly identify and understand the health care needs of this population. This study aimed to compare the differences in the prevalence of selected chronic conditions and multimorbidity, including its associated characteristics, using health survey/self-reported (SR) information only, administrative (Adm) data only and the combined (either) sources.

Methods

This was a secondary analysis of survey data from the first cycle of the Longitudinal Survey on Senior’s Health and Health Services Use linked to health-Adm data. The analytical sample consisted of 1625 community-dwelling older adults (≥65 years) recruited in the waiting rooms of primary health clinics in a selected administrative region of the province of Quebec. Seventeen chronic conditions were assessed according to two different data sources. We examined the differences in the observed prevalence of chronic conditions and multimorbidity and the agreement between data sources.

Results

The prevalence of each of the 17 chronic conditions ranged from 1.2 to 68.7% depending on the data source. The agreement between different data sources was highly variable, with kappa coefficients (κ) ranging from 0.05 to 0.73. Multimorbidity was very high in this population, with an estimated prevalence of up to 95.9%. In addition, we found that the association between sociodemographic and behavioural factors and the presence of multimorbidity varied according to the different data sources and thresholds.

Conclusions

This is the first study to simultaneously investigate chronic conditions and multimorbidity prevalence among primary care older adults using combined SR and health-Adm data. Our results call attention to (1) the possibility of underestimating cases when using a single data source and (2) the potential benefits of integrating information from different data sources to increase case identification. This is an important aspect of characterizing the health care needs of this fast-growing population.

Electronic supplementary material

The online version of this article (10.1186/s12877-019-1173-4) contains supplementary material, which is available to authorized users.

Involuntary hospitalization among young people with early psychosis: A population-based study using health administrative data

OBJECTIVE

Early psychosis is an important window for establishing long-term trajectories. Involuntary hospitalization during this period may impact subsequent service engagement in people with newly diagnosed psychotic disorder. However, population-based studies of involuntary hospitalization in early psychosis are lacking. We sought to estimate the proportion of people aged 16 to 35 years with early psychosis in Ontario who are hospitalized involuntarily at first admission, and to identify the associated risk factors and outcomes.

METHODS

Using linked population-based health administrative data, we identified incident cases of non-affective psychosis over a five-year period (2009-2013) and followed cases for two years to ascertain the first psychiatric hospitalization. We used modified Poisson regression to model sociodemographic, clinical, and service-related risk factors, and compared service-related outcomes for cases admitted on an involuntary versus voluntary basis.

RESULTS

Among 17,725 incident cases of non-affective psychosis, 38% were hospitalized within two years, and 81% of these admissions occurred on an involuntary basis (26% of cohort). Sociodemographic factors associated with an increased risk of involuntary admission included younger age (16-20), and first-generation migrant status. The strongest risk factors were poor illness insight, recent police involvement, and admission to a general (versus psychiatric) hospital. Outcomes associated with involuntary admission included increased likelihood of control intervention use and a shorter length of stay.

CONCLUSIONS

One in four young people with first-episode psychosis will have an involuntary admission early in the course of their illness. Our findings highlight areas for intervention to improve pathways to care for people with psychotic disorder.

Development and validation of a protocol for documentation of obstetric perineal lacerations

INTRODUCTION AND HYPOTHESIS

The aim of this study was to develop a new protocol for documentation of perineal lacerations and to validate the latter against the most common obstetric record system in Sweden. The hypothesis was that the new protocol would render more complete data on perineal lacerations than the current documentation method.

METHODS

A protocol for documentation of perineal lacerations was developed to be sufficiently comprehensive to serve research purposes. All women delivering their first child vaginally from 13 October 2015 to 1 February 2016 at Örebro University Hospital were eligible for the validation study. Perineal lacerations were documented using the protocol in parallel with the regular obstetric record system (ObstetriX). Cross tabulations were used to compare the coverage regarding perineal lacerations between the two documentation methods. McNemar's test was used to evaluate systematic differences between the methods.

RESULTS

A total of 187 women were included. The coverage of documentation regarding perineal laceration was significantly higher (p < 0.001) in the new protocol (89%) compared with ObstetriX (18%). Incidence of second-degree perineal tears was 26% according to the new protocol and 11% according to ObstetriX. The incidence of third-degree perineal tears A, B, and C was 2.7%, 2.1%, and 2.1%, respectively, according to the new protocol, and 3.2%, 2.7%, and 1.1% according to ObstetriX.

CONCLUSIONS

This validation study of a new documentation protocol showed that it delivered significantly more comprehensive information regarding perineal lacerations than the most common obstetric record system in Sweden.

Unmet need for mental health services among people screened but not admitted to an early psychosis intervention program

Information is lacking on people screened for early psychosis intervention (EPI) but not admitted to the program. Using health administrative data, we constructed a retrospective cohort of incident cases of psychosis in the catchment of an EPI program. Use of mental health services was compared between people screened and not admitted with an EPI-admitted group. The non-admitted group had higher rates of subsequent emergency department visits, psychiatric hospitalizations, and involuntary admissions. These patterns are indicative of unmet need, and people screened but not admitted to EPI may benefit from protocols to improve transitions of care with other service providers.

Predictive models of disease burden at diagnosis in persons with adult-onset ulcerative colitis using health administrative data

Background

Health administrative data is increasingly used to conduct population-based health services research. A major limitation of these data for the study of inflammatory bowel diseases is the absence of detailed clinical information relating to disease burden. We used Ontario health administrative data to develop predictive models of disease burden at diagnosis in ulcerative colitis (UC) patients for future use in population-based studies of incident UC cohorts.

Methods

Through chart review, we characterized macroscopic colitis activity and extent at diagnosis in consecutive adult-onset UC patients diagnosed at The Ottawa Hospital between 2001 and 2012. We linked this cohort to Ontario health administrative data to test the capacity of administrative variables to discriminate different levels of disease activity, disease extent and the disease burden (a composite of disease extent and activity). We modelled outcomes as binary (using logistic regression) and ordinal (using proportional odds regression) variables and performed bootstrap validation of our final models.

Results

We tested 20 administrative variables in 587 eligible patients. The logistic model of total disease burden (severe and extensive colitis vs. all other phenotypes) showed moderate discriminatory capacity (optimism-corrected c-statistic value 0.729). Individual models of disease extent and disease activity showed poorer discriminatory capacity (c-statistic value < 0.7 for 3 of 4 models).

Conclusions

Ontario health administrative data may reasonably discriminate levels of total disease burden at diagnosis in adult-onset UC patients. Our models should be externally validated before their widespread application in future population-based studies of incident UC cohorts to adjust for the confounding effects of differences in disease burden.

Electronic supplementary material

The online version of this article (10.1186/s12876-018-0924-6) contains supplementary material, which is available to authorized users.

Persistence and adherence to the new oral disease-modifying therapies for multiple sclerosis: A population-based study

OBJECTIVE

To examine persistence and adherence to the oral disease-modifying therapies (DMTs) for multiple sclerosis (MS).

METHODS

Population-based health administrative databases in British Columbia, Canada were accessed to identify all individuals filling an oral DMT prescription for MS (fingolimod, dimethyl fumarate, teriflunomide) between January 2011 and December 2015. Predictors of persistence and adherence at 6 and 12 months were assessed using logistic regression, with estimates expressed as adjusted odds ratios (aORs), and 95% confidence intervals (CIs).

RESULTS

Of 858 individuals with ≥6 months of follow-up, the mean age at first prescription was 43.0 (SD:10.3) years; 74.2% were women. By 6 months 11.0% (94/858) had discontinued their initial oral DMT; by 12 months the proportion was 19.6% (113/577). Over 6 and 12 months, among those persisting with their oral DMT, 82.5% (630/764) and 81.7% (379/464) exhibited optimal adherence (proportion of days covered ≥80%). Age, sex, calendar year and comorbidity were not associated with persistence or adherence. Individuals with higher neighbourhood-level socioeconomic status had higher odds of discontinuation within 6 months (aOR = 2.2; 95%CI:1.3-3.7). Those who had previously used another DMT had higher odds of optimal adherence (6 months aOR = 2.4;95%CI:1.6-3.6, and 12 months aOR = 2.4; 95%CI:1.5-3.9).

CONCLUSION

Approximately 1 in 10 individuals discontinued their first oral DMT within 6 months, and 1 in 5 did so within one year. However, among those who did continue drug, a high proportion (>80%) exhibited optimal adherence. Predictors of persistence or adherence with immediate practical application were lacking; this highlights the challenges in optimizing drug therapy.

The burden of infant group B streptococcal infections in Ontario: Analysis of administrative data to estimate the potential benefits of new vaccines

Group B streptococcus (GBS) is a leading bacterial cause of neonatal sepsis and meningitis in many countries as well as an important cause of disease in pregnant women. Currently, serotype-specific conjugate vaccines are being developed. We conducted an epidemiological analysis of health administrative data to estimate the burden of infant GBS disease in Ontario, Canada and combined these estimates with literature on serotype distribution to estimate the burden of disease likely to be vaccine-preventable. Between 1st January 2005 and 31st December 2015, 907 of 64320 health care encounters in Ontario in patients under 1 year old had codes specifically identifying GBS as the cause of the disease, of which 717 were under one month of age. In addition, application of epidemiological data to the remaining patients allowed us to estimate a further 2322 cases and among them 1822 were under one month of age. In the same period, 579 confirmed neonatal invasive GBS cases in patients up to one month of age were reported to public health. Depending on serotype distribution, vaccination coverage and early versus late onset disease (0-6 days and 7-90 days of age respectively), the preventable fraction ranged widely. With a vaccine that is 90% effective and 60% immunization coverage, up to 52% of early and late onset disease could be prevented by forthcoming vaccines. GBS is under-reported in Ontario. Uncertainty about the potential impact of vaccine indicates that further analysis and research may be needed to prepare for policy-decision making, including clinical validation studies and an economic evaluation of GBS vaccination in Ontario.

Informing Targeted Interventions to Optimize the Cascade of HIV Care Using Cluster Analyses of Health Resource Use Among People Living with HIV/AIDS

Increased awareness of the secondary preventive benefits of antiretroviral treatment (ART) has strengthened the desire to optimize health care systems' response to HIV/AIDS. We identified clusters of health resource utilization (HRU) among people living with HIV (PLHIV) to inform targeted interventions aimed to optimize the cascade of HIV care. Using linked population-level health databases in British Columbia, Canada, we selected two analytic samples of PLHIV with 3 years of follow-up between 2006-2011 that were classified as intermittently retained in care or intermittently engaged in ART, and executed a probabilistic model-based clustering analysis for each sample with 5 and 9 quarterly HRU variables, respectively. We found clear HRU profile differences among both samples with similar HIV-related care: one featured active involvement in non-HIV care, the other little or no health care interaction following linkage to care. Differential reengagement intervention strategies capitalizing on missed opportunities in non-HIV care and further engaging physicians delivering HIV care are needed to optimize the response to the HIV epidemic.

Prevaccination epidemiology of herpes zoster in Denmark: Quantification of occurrence and risk factors

BACKGROUND

Herpes zoster (HZ) is a vaccine-preventable disease caused by reactivation of the varicella-zoster virus. Unfortunately, formulation of recommendations on routine immunization is hampered by a lack of data on disease burden, since most countries do not record cases of HZ in the general population. We developed and validated an algorithm to identify HZ based on routinely collected registry data and used it to quantify HZ occurrence and risk factors in Denmark prior to marketing of the HZ vaccine.

METHODS

We included patients aged ≥40years with a first-time systemic Acyclovir, Valacyclovir, or Famciclovir prescription or a hospital-based HZ diagnosis in the Danish nationwide health registries during 1997-2013. In a validation substudy (n=176), we computed the proportion of persons with HZ among patients who redeemed antiviral prescriptions. In a cohort study, we computed age-specific rates of HZ (45,297,258 person-years). In a case-control study, we then computed odds ratios (ORs) for common chronic diseases and immunosuppressive factors among HZ cases (n=189,025) vs. matched population controls (n=945,111).

RESULTS

Medical record review confirmed HZ in 87% (95% confidence interval: 79-93%) of persons ≥40years who dispensed antivirals at doses recommended for HZ. HZ rates increased from 2.15/1000 person-years in 40-year-olds to 9.45/1000 person-years in 95-year-olds. Rates were highest in women. HZ was diagnosed during hospitalization among 3.5%. As expected, persons with severe immunosuppressive conditions had the highest ORs of HZ (between 1.82 and 4.12), but various autoimmune diseases, asthma, chronic kidney disease, and inhaled glucocorticoids were also associated with increased ORs (between 1.06 and 1.64).

CONCLUSION

This algorithm is a valid tool for identifying HZ in routine healthcare data. It shows that HZ is common in Denmark, especially in patients with certain chronic conditions. Prioritized vaccination of such high-risk patients might be an option in countries considering alternatives to universal vaccination.

A cohort study examining emergency department visits and hospital admissions among people who use drugs in Ottawa, Canada

Background

The health of people who use drugs (PWUD) is characterized by multimorbidity and chronicity of health conditions, necessitating an understanding of their health care utilization. The objective of this study was to evaluate emergency department (ED) visits and hospital admissions among a cohort of PWUD.

Methods

We used a retrospective observational design between 2012 and 2013. The population was a marginalized cohort of PWUD (the PROUD study) for whom survey data was linked (n = 663) to provincial health administrative data housed at the Institute for Clinical Evaluative Sciences. We constructed a 5:1 comparison group matched by age, sex, income quintile, and region. The main outcomes were defined as having two or more ED visits, or one or more hospital admissions, in the year prior to survey completion. We used multivariable logistic regression analyses to identify factors associated with these outcomes.

Results

Compared to the matched cohort, PWUD had higher rates of ED visits (rate ratio [RR] 7.0; 95% confidence interval [95% CI] 6.5–7.6) and hospitalization (RR 7.7; 95% CI 5.9–10.0). After adjustment, factors predicting more ED visits were receiving disability (adjusted odds ratio [AOR] 3.0; 95% CI 1.7–5.5) or income assistance (AOR 2.7; 95% CI 1.5–5.0), injection drug use (AOR 2.1; 95% CI 1.3–3.4), incarceration within 12 months (AOR 1.6; 95% CI 1.1–2.4), mental health comorbidity (AOR 2.1; 95% CI 1.4–3.1), and a suicide attempt within 12 months (AOR 2.1; 95% CI 1.1–3.4). Receiving methadone (AOR 0.5; 95% CI 0.3–0.9) and having a regular family physician (AOR 0.5; 95% CI 0.2–0.9) were associated with lower odds of having more ED visits. Factors associated with more hospital admissions included Aboriginal identity (AOR 2.4; 95% CI 1.4–4.1), receiving disability (AOR 2.4; 95% CI 1.1–5.4), non-injection drug use (opioids and non-opioids) (AOR 2.2; 95% CI 1.1–4.4), comorbid HIV (AOR 2.4; 95% CI 1.2–5.6), mental health comorbidity (AOR 2.4; 95% CI 1.3–4.2), and unstable housing (AOR 1.9; 95% CI 1.0–3.4); there were no protective factors for hospitalization.

Conclusions

Improved post-incarceration support, housing services, and access to integrated primary care services including opioid replacement therapy may be effective interventions to decrease acute care use among PWUD, including targeted approaches for people receiving social assistance or with mental health concerns.

Bootstrap imputation with a disease probability model minimized bias from misclassification due to administrative database codes

OBJECTIVE

Diagnostic codes used in administrative databases cause bias due to misclassification of patient disease status. It is unclear which methods minimize this bias.

STUDY DESIGN AND SETTING

Serum creatinine measures were used to determine severe renal failure status in 50,074 hospitalized patients. The true prevalence of severe renal failure and its association with covariates were measured. These were compared to results for which renal failure status was determined using surrogate measures including the following: (1) diagnostic codes; (2) categorization of probability estimates of renal failure determined from a previously validated model; or (3) bootstrap methods imputation of disease status using model-derived probability estimates.

RESULTS

Bias in estimates of severe renal failure prevalence and its association with covariates were minimal when bootstrap methods were used to impute renal failure status from model-based probability estimates. In contrast, biases were extensive when renal failure status was determined using codes or methods in which model-based condition probability was categorized.

CONCLUSION

Bias due to misclassification from inaccurate diagnostic codes can be minimized using bootstrap methods to impute condition status using multivariable model-derived probability estimates.

Prevalence of multiple sclerosis in the Lazio region, Italy: use of an algorithm based on health information systems

Compared with other areas of the country, very limited data are available on multiple sclerosis (MS) prevalence in Central Italy. We aimed to estimate MS prevalence in the Lazio region and its geographical distribution using regional health information systems (HIS). To identify MS cases we used data from drug prescription, hospital discharge and ticket exemption registries. Crude, age- and gender-specific prevalence estimates on December 31, 2011 were calculated. To compare MS prevalence between different areas within the region, we calculated age- and gender-adjusted prevalence and prevalence ratios using a multivariate Poisson regression model. Crude prevalence rate was 130.5/100,000 (95 % CI 127.5–133.5): 89.7/100,000 for males and 167.9/100,000 for females. The overall prevalence rate standardized to the European Standard Population was 119.6/100,000 (95 % CI 116.8–122.4). We observed significant differences in MS prevalence within the region, with estimates ranging from 96.3 (95 % CI 86.4–107.3) for Latina to 169.6 (95 % CI 147.6–194.9) for Rieti. Most districts close to the coast showed lower prevalence estimates compared to those situated in the eastern mountainous area of the region. In conclusion, this study produced a MS prevalence estimate at regional level using population-based health administrative databases. Our results showed the Lazio region is a high-risk area for MS, although with an uneven geographical distribution. While some limitations must be considered including possible prevalence underestimation, HIS represent a valuable source of information to measure the burden of SM, useful for epidemiological surveillance and healthcare planning.

International variation in medication prescription rates among elderly patients with inflammatory bowel disease

BACKGROUND AND AIMS

The elderly represent a growing demographic of patients with IBD. No study has previously described variations in care or medication prescriptions in senior patients with IBD. We compared prescription rates among elderly patients with IBD in four countries using health administrative data.

METHODS

Databases from the United States (US), United Kingdom (UK), Denmark and Canada were queried. Variation in prescription rates between countries was assessed in patients ≥65y with prevalent IBD who had ≥1 prescription for an IBD-related medication in a given quarter between 2004 and 2009. Patients were identified using previously-reported, validated algorithms. Country-specific rates were compared in each quarter using Fisher's exact test.

RESULTS

In patients with Crohn's disease, Canada and US had higher prescription rates for oral 5-ASA (P<0.0001 in all quarters) and infliximab (P<0.05 in 22/24 quarters), while the US had higher rates of thiopurine usage (P<0.05 in 23/24 quarters). Canada had greater rates of methotrexate prescriptions (P<0.05 in 21/24 quarters analyzed). In patients with ulcerative colitis (UC), rates of oral steroid usage was lowest in the US (P<0.05 in 22/24 quarters) and oral 5-ASA use was highest in the US and Canada (P<0.0001 in all quarters). Canada and Denmark used more rectal therapy than the US. Infliximab usage in UC was significantly higher in the US and Canada after 2006.

CONCLUSIONS

Significant variation in medication prescription rates exists among countries. Future research should assess whether these differences were associated with disparities in outcomes and health care costs.