MATCHING AND INTEGRATION OF REGISTRY AND SURVEY DATA ON THE DYNAMICS OF WORK HISTORIES: A PILOT STUDY

EnglishThe paper deals with changes in the work histories' patterns of two cohorts of young people living in the province of Trento (Italy) and aged 15-29 years, who got their first job in 1974/95 and 1999/2011, respectively. We build and use a database resulting from the deterministic matching and integration of two datasets: a multi-purpose panel survey carried out on a representative sample of households, and register data from the provincial section of the administrative archive of INPS -the Italian social security agency. From a methodological point of view, the integrated database gives clear evidence that register data provide much more detailed and reliable information on work histories than survey data. From a substantive perspective, we find sizeable changes across the two cohorts in (i) first job spells, (ii) number and length of work episodes in a time-window of eight years from entry in the first job, and (iii) the overall pattern of job participation and job mobility during the initial eight years of work histories. On the contrary, no variation across cohorts in career mobility patterns is observed. The evidence consistently points to a significant worsening of the working experience of the youth over less than two decades. With appropriate caveats, this evidence appears to be extensible to northern-central Italy. italianoIntroduzione. - L'esperienza pilota di integrazione e il disegno di formazione delle coorti. - Il primo lavoro. - L'occupazione nei primi otto anni di storia lavorativa: un'analisi centrata sugli episodi. - L'occupazione nei primi otto anni di storia lavorativa: un'analisi centrata sulle storie lavorative. - La mobilita di carriera. - Riflessioni conclusive.

In this chapter, our focus is on panel conditioning with respect to attitude questions. Our methodological approach is different from the majority of previous studies in this area, in that we do not attempt to estimate biases in marginal and associational distributions through comparison with a fresh cross-sectional sample. Rather, our approach is based on testing hypotheses on a single dataset, derived from an explicit theoretical model of the psychological mechanism underlying conditioning effects in repeated measures of the attitude. We refer to this theoretical account as the cognitive stimulus (CS) model. Specifically, we use a range of empirical indicators to evaluate the theory that repeatedly administering attitude questions serves to stimulate respondents to reflect and deliberate more closely on the issues to which the questions pertain. This, in turn, results in stronger and more internally consistent attitudes in the later waves of a panel. The chapter proceeds in the following manner. First, we review the existing literature on panel conditioning effects. Next, we set out in more detail the rationale underlying the CS hypothesis.We then use data from the first ten waves of the British Household Panel Study (BHPS) to test four inter-related hypotheses expressed as empirical expectations of the CS model. We conclude with a discussion of the implications of our findings for the validity of attitude measures in panel surveys.

To characterize agreement between administrative and registry data in the determination of patient-level comorbidities. Previous research finds poor agreement between these 2 types of data in the determination of outcomes. We hypothesized that concordance between administrative and registry data would also be poor. A cohort of inpatient operations (length of stay 1 day or greater) was obtained from a consortium of 8 hospitals. Within each hospital, National Surgical Quality Improvement Program (NSQIP) data were merged with intra-institutional inpatient administrative data. Twelve different comorbidities (diabetes, hypertension, congestive heart failure, hemodialysis-dependence, cancer diagnosis, chronic obstructive pulmonary disease, ascites, sepsis, smoking, steroid, congestive heart failure, acute renal failure, and dyspnea) were analyzed in terms of agreement between administrative and NSQIP data. Forty-one thousand four hundred thirty-two inpatient surgical hospitalizations were analyzed in this study. Concordance (Cohen Kappa value) between the 2 data sources varied from 0.79 (diabetes) to 0.02 (dyspnea). Hospital variation in concordance (intersite variation) was quantified using a test of homogeneity. This test found significant intersite variation at a level of P < 0.001 for each of the comorbidities except for dialysis (P = 0.07) and acute renal failure (P = 0.19). These findings imply significant differences between hospitals in their generation of comorbidity data. This study finds significant differences in how administrative versus registry data assess patient-level comorbidity. These differences are of concern to patients, payers, and providers, each of which had a stake in the integrity of these data. Standardized definitions of comorbidity and periodic audits are necessary to ensure data accuracy and minimize bias.

The past decade has seen a dramatic increase in efforts by both payors and governmental agencies to measure and publicly report patient outcomes to ensure that patients receive appropriate, high-quality care. As a result, terms such as outcomes, quality, report cards, and benchmarking have entered the lexicon of clinical medicine, and both physicians and hospitals are increasingly being held accountable for the outcomes of their patients. This movement shows no sign of slowing down. In May 2008, the Centers for Medicare and Medicaid Services proposed increasing the total number of inpatient hospital measures from 40 to 73,1 and numerous physician measures are in various phases of development. In the current environment, it is likely that implantable cardioverter defibrillators (ICDs), which reduce the risk of sudden cardiac death,2 will soon emerge as a high-priority target for the development of performance measures, given that the procedure is expensive, performed on severely ill patients, and associated with significant risk of complications. Article see p 240 Accordingly, the findings published by Al-Khatib et al3 in this issue of Circulation: Arrhythmia and Electrophysiology warrant consideration in the context of current efforts to publicly report patient outcomes using Medicare claims data. In this excellent article, the authors used data from Centers for Medicare and Medicaid Services standard analytic files to determine the 90-day postprocedure complication rates and 1-year mortality rates of 8581 Medicare beneficiaries who had an ICD implanted by one of 1959 implanting physicians. The authors note that the complication rate declined from 18.8% in 2002 to 14.2% in 2005 and further identified specific factors associated with an increased risk of complication such as chronic lung disease, dementia, and renal disease. Of interest, physician experience as measured by procedure volume was not significantly associated with the risk of complication, although an earlier …

This paper shows how Danish administrative register data can be combined with survey data at the person level and be used to validate information collected in the survey. Register data are collected by automatic third party reporting and the potential errors associated with the two data sources are therefore plausibly orthogonal. Two examples are given to illustrate the potential of combining survey and register data. In the first example expenditure survey records with information about total expenditure are merged with income tax records holding information about income and wealth. Income and wealth data are used to impute total expenditure which is then compared to the survey measure. Results suggest that the two measures match each other well on average. In the second example we compare responses to a one-shot recall question about total gross personal income (collected in another survey) with tax records. Tax records hold detailed information about different types of income and this makes it possible to test if errors in the survey response are related to the reporting of particular types of income. Results show bias in the mean and that the survey error has substantial variance. Results also show that the errors are correlated with conventional covariates suggesting that the errors are not of the classical type. The latter example illustrates how Denmark can be used as a "laboratory" for future validation studies. Tax records with detailed information about different types of income are available for the entire Danish population and can be readily merged to survey data. This makes it possible to test the ability of respondents to accurately report different types of income using different interviewing techniques and questions. The examples presented in this paper are based on cross section data. However, the possibility to issue surveys repeatedly to the same persons and linking up to longitudinal tax records provides an opportunity to learn more about the time series properties of measurement errors, a subject about which little evidence exist, in the future.

Background Population level information on the prevalence of hypertension and diabetes is needed to support planning and evaluation of preventive activities and care. Population based health examination surveys (HES) can provide valid information, but they are time consuming and expensive. Administrative registers on hospitalizations and out-patient visits could provide information faster and at less cost. The aim of this study is to estimate how comparable prevalence estimates of type 2 diabetes (T2D), hypertension and their combination are based on data from the Finnish national health examination survey (FinHealth 2017) and data from administrative hospital and primary care registers in Finland. Methods Survey data were linked to care registers using personal identity codes. Survey based hypertension was defined as SBP≥140 mmHg or DBP≥90 mmHg or self-reported use of antihypertensive medications; and T2D as HbA1c ≥48 mmol/mol or self-reported us of diabetes medications. Corresponding indicators from care registers were: hypertension and T2D as a reason for care or visit or having received prescription for diabetes medication (using ICD-10, ICPC and ATC codes). Results For hypertension, survey data provided the prevalence of 43 % while only 12 % of individuals were identified as hypertensives in the register data. The prevalence of T2D was 9 % in both data sources. The prevalence of having both hypertension and T2D was 7 % based on survey data and 3 % based on register data. Agreement between survey and register data was lower for hypertension (Cohen’s kappa 0.23) than for T2D (0.84). Conclusions Register data provided lower prevalence for hypertension than the survey data. For diabetes, similar prevalences were observed. As there are limitations in the coverage of register data more reliable population level information can be obtained from HES. Key messages Health examination surveys cover persons with undiagnosed problems and conditions omitted in national registers. Hypertension and T2D should be monitored with both register and survey data.

Quantitative research on volunteering often relies upon either cross-sectional survey data or general interest longitudinal surveys with simple one-item questions about volunteering. In countries such as the Scandinavian where administrative register data are readily available for research purposes, an enriching survey data with administrative register data offers researchers the opportunity to gain a richer understanding of voluntary action. This chapter discusses the benefits and challenges of merging longitudinal survey data on volunteering with administrative register data at the individual level in Denmark. These benefits and challenges are explored by drawing on recent empirical examples from research on trends in volunteering over time, immigrant volunteering, and the effect of volunteering on wages. I conclude the chapter by summarizing the benefits of merging longitudinal survey data with data from administrative registers for volunteerism research and discuss possible limitations.

BackgroundAdministrative and survey data are two key data sources for population-based research about chronic disease. The objectives of this methodological paper are to: (1) estimate agreement between the two data sources for irritable bowel syndrome (IBS) and compare the results to those for inflammatory bowel disease (IBD); (2) compare the frequency of IBS-related diagnoses in administrative data for survey respondents with and without self-reported IBS, and (3) estimate IBS prevalence from both sources.MethodsThis retrospective cohort study used linked administrative and health survey data for 5,134 adults from the province of Manitoba, Canada. Diagnoses in hospital and physician administrative data were investigated for respondents with self-reported IBS, IBD, and no bowel disorder. Agreement between survey and administrative data was estimated using the κ statistic. The χ2 statistic tested the association between the frequency of IBS-related diagnoses and self-reported IBS. Crude, sex-specific, and age-specific IBS prevalence estimates were calculated from both sources.ResultsOverall, 3.0% of the cohort had self-reported IBS, 0.8% had self-reported IBD, and 95.3% reported no bowel disorder. Agreement was poor to fair for IBS and substantially higher for IBD. The most frequent IBS-related diagnoses among the cohort were anxiety disorders (34.4%), symptoms of the abdomen and pelvis (26.9%), and diverticulitis of the intestine (10.6%). Crude IBS prevalence estimates from both sources were lower than those reported previously.ConclusionsPoor agreement between administrative and survey data for IBS may account for differences in the results of health services and outcomes research using these sources. Further research is needed to identify the optimal method(s) to ascertain IBS cases in both data sources.

BackgroundAdministrative health data and cardiac device registries can be used to empirically evaluate outcomes and costs after implantable cardioverter defibrillator (ICD) implantation. These datasets often have incomplete information on the indication for implantation (primary vs secondary prevention of sudden cardiac death). MethodsWe used 16 years of population-based cardiac device registry and administrative health data from British Columbia, Canada, to derive and internally validate statistical models that predict the likely indication for ICD implantation. We used chart review data as the reference standard for ICD indication in the Cardiac Device Registry database (CDR; 2004-2012 [Cardiac Services BC]) and nonmissing indication as the reference standard in the Heart Information System registry database (HEARTis; 2013-2019 [Cardiac Services BC]). We created 3 logistic regression prediction models in each database: one using only registry data, one using only administrative data, and one using both registry and administrative data. We assessed the predictive performance of each model using standard metrics after optimism correction with 200 bootstrap resamples. ResultsModels that used registry data alone demonstrated excellent predictive performance (sensitivity ≥ 89%; specificity ≥ 87%). Models that used only administrative data performed well (sensitivity ≥ 84%; specificity ≥ 70%). Models that used both registry and administrative data showed modest gains over those that used registry data alone (sensitivity ≥ 90%; specificity ≥ 89%). ConclusionsAdministrative health data and cardiac device registry data can distinguish secondary prevention ICDs from primary prevention ICDs with acceptable sensitivity and specificity. Imputation of missing ICD indication might make these data resources more useful for research and health system monitoring.

Background: The California (CA) Office of Statewide Health Planning and Development (OSHPD) collects three types of patient-level administrative data: patient discharge (PDD), ambulatory surgery (AS), and emergency room (ER). CA hospitals are mandated to submit clinical data to OSHPD for all adult coronary artery bypass graft (CABG) surgeries. This submission is usually a subset of hospital’s Society of Thoracic Surgeons Adult Cardiac Surgery Database submissions. In addition, OSHPD receives clinical data from the National Cardiac Data Registry CathPCI for CA hospitals certified to perform elective PCIs without onsite surgery available. OSHPD implements various data quality efforts to ensure accuracy in each data source, including using data from administrative registries to confirm clinical data and using data from clinical registries to confirm administrative data. Methods: The clinical CABG data submitted to OSHPD is linked to the PDD. The resulting discrepancy reports are shared with hospitals to assist them when revising their data. The reports show discrepancies in over- and under-reporting of CABG cases, isolated and non-isolated CABGs, and post-op complications. For PCI reporting, clinical data from the CathPCI registry is compared to the PDD, AS, and ER to look for discrepancies. Results: For the 2016 CABG data, discrepancy reports issued to 126 hospitals submitting clinical CABG data to OSHPD showed potential under-reporting of 417 CABGs that appeared in the PDD but were not submitted in the clinical data. The report showed an additional 437 CABGs that were submitted in the clinical data, but did not appear in the PDD. Hospital review and response to these reports resulted in an overall increase of 296 CABGs to the clinical database. Discrepancy reports also alerted hospitals to 66 cases where a post-operative stroke occurred but was not reported in the clinical database. Hospital review resulted in 22 post-operative strokes added to the clinical registry. The 2016 CathPCI data for CA certified elective PCI hospitals showed potential under reporting of PCIs in PDD, AS, and ER. Possible missing PCIs ranged from 0.4% to 21.6% per hospital. OSHPD worked with staff at the clinical data and administrative data units at these hospitals to understand the discrepancies. The main issue identified was that AS data submitted by some hospitals used alternate coding (Healthcare Common Procedure Coding System) and failed to convert these codes to Current Procedural Terminology codes as required. The findings plan to be used in an outreach effort to all CA hospital submitting administrative PCI data to OSHPD to ensure accurate reporting. Conclusion: Comparing administrative and clinical data registries data is an effective quality tool to identify discrepancies in each source of data. Sharing discrepancies with hospitals results in improved understanding of data standards and data quality.

Impact evaluations draw their data from two sources, namely, surveys conducted for the evaluation or administrative data collected for other purposes. Both types of data have been used in impact evaluations of social programs. This study analyzes the causes of differences in impact estimates when survey data and administrative data are used to evaluate earnings impacts in social experiments and discusses the differences observed in eight evaluations of social experiments that used both survey and administrative data. There are important trade-offs between the two data sources. Administrative data are less expensive but may not cover all income and may not cover the time period desired, while surveys can be designed to avoid these problems. We note that errors can be due to nonresponse or reporting, and errors can be balanced between the treatment and the control groups or unbalanced. We find that earnings are usually higher in survey data than in administrative data due to differences in coverage and likely overreporting of overtime hours and pay in survey data. Evaluations using survey data usually find greater impacts, sometimes much greater. The much lower cost of administrative data make their use attractive, but they are still subject to underreporting and other problems. We recommend further evaluations using both types of data with investigative audits to better understand the sources and magnitudes of errors in both survey and administrative data so that appropriate corrections to the data can be made.

The rocky exhilarating journey from data to wisdom

BackgroundNEPS-SC6-ADIAB is a new linked data product containing survey data of Starting Cohort 6 of the German National Educational Panel Study (NEPS) and administrative employment data from the Institute for Employment Research (IAB), the research institute of the Federal Employment Agency. NEPS is provided by the Leibniz Institute for Educational Trajectories (LIfBi). Starting Cohort 6 of this panel survey includes adults in their professional life, the survey focuses on education in adulthood and lifelong learning. The administrative data in NEPS-SC6-ADIAB consist of comprehensive information on the employment histories.&#x0D; ObjectivesCombining these two data sources increases for example the information about individual employment history. Overall, the data volume is increased by the linkage between the survey data and the administrative data.&#x0D; MethodsA record linkage process was used to link the two data sources. The data access is free for the whole scientific community. In addition to a large number of On-site access locations within Germany, there are also international On-site access locations. Including London and Colchester. In addition a Remote Data Access is offered.&#x0D; ConclusionsThis data linkage project is very innovative and creates an extensive database, which results in extensive analytical potential. A short application example is made to exemplify the comprehensive analytical potential of NEPS-SC6-ADIAB. This ongoing project deals with nonresponse in survey data. The linked data has a variety of variables collected in both data sources, administratively and through the NEPS survey, allowing for comparative analyses. In this case an idea to compensate nonresponse in income data with administrative data is drawn.

BackgroundAdministrative health care databases may be particularly useful for injury surveillance, given that they are population-based, readily available, and relatively complete. Surveillance based on administrative data, though, is often restricted to injuries that result in hospitalization. Adding physician billing data to administrative data-based surveillance efforts may improve comprehensiveness, but the feasibility of such an approach has rarely been examined. It is also not clear how injury surveillance information obtained using administrative health care databases compares with that obtained using self-report surveys. This study explored the value of using physician billing data along with hospitalization data for the surveillance of adolescent injuries in Ontario, Canada. We aimed i) to document the burden of adolescent injury using administrative health care data, focusing on the relative contribution of physician billing information; and ii) to explore data quality issues by directly comparing adolescent injuries identified in administrative and self-report data.MethodsThe sample included adolescents aged 12 to 19 years who participated in the 1996–1997 cross-sectional Ontario Health Survey, and whose survey responses were linked to administrative health care datasets (N = 2067). Descriptive analysis was used to document the burden of injuries as a proportion of all physician care by gender and location of care, and to examine the distribution of both administratively-defined and self-reported activity-limiting injuries according to demographic characteristics. Administratively-defined and self-reported injuries were also directly compared at the individual level.ResultsApproximately 10% of physician care for the sample was identified as injury-related. While 18.8% of adolescents had self-reported injury in the previous year, 25.0% had documented administratively-defined injury. The distribution of injuries according to demographic characteristics was similar across data sources, but congruence was low at the individual level. Possible reasons for discrepancies between the data sources included recall errors in the survey data and errors in the physician billing data algorithm.ConclusionIf further validated, physician billing data could be used along with hospital inpatient data to make an important and unique contribution to adolescent injury surveillance. The limitations inherent in different datasets highlight the need to continue rely on multiple information sources for complete injury surveillance information.

Administrative Data Sets are Inaccurate for Assessing Functional Outcomes After Radical Prostatectomy