Differentiating Incident from Recurrent Stroke Using Administrative Data: The Impact of Varying Lengths of Look-Back Periods on the Risk of Misclassification

Impact of observability period on the classification of COPD diagnosis timing among Medicare beneficiaries with lung cancer

BACKGROUND

Investigators often use claims data to estimate the diagnosis timing of chronic conditions. However, misclassification of chronic conditions is common due to variability in healthcare utilization and in claims history across patients.

OBJECTIVE

We aimed to quantify the effect of various Medicare fee-for-service continuous enrollment period and lookback period (LBP) on misclassification of COPD and sample size.

METHODS

A stepwise tutorial to classify COPD, based on its diagnosis timing relative to lung cancer diagnosis using the Surveillance Epidemiology and End Results cancer registry linked to Medicare insurance claims. We used 3 approaches varying the LBP and required continuous enrollment (i.e., observability) period between 1 to 5 years. Patients with lung cancer were classified based on their COPD related healthcare utilization into 3 groups: pre-existing COPD (diagnosis at least 3 months before lung cancer diagnosis), concurrent COPD (diagnosis during the -/+ 3months of lung cancer diagnosis), and non-COPD. Among those with 5 years of continuous enrollment, we estimated the sensitivity of the LBP to ascertain COPD diagnosis as the number of patients with pre-existing COPD using a shorter LBP divided by the number of patients with pre-existing COPD using a longer LBP.

RESULTS

Extending the LBP from 1 to 5 years increased prevalence of pre-existing COPD from ~ 36% to 51%, decreased both concurrent COPD from ~ 34% to 23% and non-COPD from ~ 29% to 25%. There was minimal effect of extending the required continuous enrollment period beyond one year across various LBPs. In those with 5 years of continuous enrollment, sensitivity of COPD classification (95% CI) increased with longer LBP from 70.1% (69.7% to 70.4%) for one-year LBP to 100% for 5-years LBP.

CONCLUSION

The length of optimum LBP and continuous enrollment period depends on the context of the research question and the data generating mechanisms. Among Medicare beneficiaries, the best approach to identify diagnosis timing of COPD relative to lung cancer diagnosis is to use all available LBP with at least one year of required continuous enrollment.

Predictors of recurrent stroke and subsequent mortality among patients experiencing housing instability

BACKGROUND

The experience of homelessness has been linked with developing poor health outcomes. Little is known about the risk of recurrent stroke among these individuals. This study investigated the correlates of developing recurrent stroke and subsequent mortality among Veterans with housing instability.

METHODS

Using a national sample of Veterans from the U.S. Department of Veterans Affairs who had an indicator of housing instability between 2014-2018 (n=659,987), we identified 15,566 Veterans who experienced incident stroke. We compared characteristics of Veterans who experienced incident stroke and did and did not experience recurrent stroke and conducted logistic regressions using a discrete-time survival framework to assess two outcomes: recurrent stroke and all-cause mortality.

RESULTS

Among our cohort, 91.3% did not experience recurrent stroke while 8.7% did during the observation period. The receipt of any level of primary care outpatient visits was associated with a reduction in the odds of recurrent stroke. Several medical diagnoses were also associated with increased odds of recurrent stroke, including hypertension (aOR 1.35, 95% CI 1.15-1.59), diabetes (aOR 1.21, 95% CI 1.07-1.36), and renal disease (aOR 1.17, 95% CI 1.02, 1.35). Veterans who used any level of VA Homeless Programs had reduced odds of all-cause mortality (high level: aOR 0.65, 95% CI 0.60-0.71; low level: aOR 0.66, 95% CI 0.60-0.73).

CONCLUSION

Our study found several predictors of developing recurrent stroke and subsequent death in a population of Veterans experiencing housing instability. Implications include the need to monitor closely high-risk patients who have experienced incident stroke and have other co-occurring needs.

Evaluation of Left Truncation and Censoring When Changing the Use of the International Classification of Diseases Eighth Revision Codes to Tenth Revision Codes in the Danish National Patient Registry

Purpose

In the Danish National Patient Registry (DNPR), covering all Danish hospitals and widely used in research, diseases have been recorded using International Classification of Diseases (ICD) codes, transitioning from the Eighth to the Tenth revision in 1994. Uncertainty exists regarding whether including ICD-8 codes alongside ICD-10 is needed for complete disease identification. We assessed the extent of left-truncation and left-censoring in the DNPR arising from omitting ICD-8 codes.

Patients and Methods

We sampled 500,000 Danes ≥40 years of age in 1995, 2010, and 2018. From the DNPR, we identified cardiovascular, endocrine, gastrointestinal, neurological, pulmonary, rheumatic, and urogenital diseases as well as fractures. We obtained the number of people with a disease recorded with ICD-8 codes only (ie, the ICD-8 record would be left-truncated by not using ICD-8 codes), ICD-8 plus ICD-10 codes (ie, the ICD-8 record would be left-censored by not using ICD-8 codes), and ICD-10 codes only. For each ICD group, we calculated the proportion of people with the disease relative to the total sample (ie, 500,000 people) and the total number of people with the disease across all ICD groups.

Results

Overall, the left-truncation issue decreased over the years. Relative to all people with a disease, the left-truncated proportion was for example 59% in 1995 and <2% in 2018 for diabetes mellitus; 93% in 1995, and 54% in 2018 for appendicitis. The left-truncation issue increased with age group for most diseases. The proportion of disease records left-censored by not using ICD-8 codes was generally low but highest for chronic diseases.

Conclusion

The left-truncation issue diminished over sample years, particularly for chronic diseases, yet remained rather high for selected surgical diseases. The left-truncation issue increased with age group for most diseases. Left-censoring was overall a minor issue that primarily concerned chronic diseases.

Socioeconomic determinants of myalgic encephalomyelitis/chronic fatigue syndrome in Norway: a registry study

BACKGROUND

Previous research has shown that socioeconomic status (SES) is a strong predictor of chronic disease. However, to the best of our knowledge, there has been no studies of how SES affects the risk of Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) that has not been based upon self-reporting or retrospectively screening of symptoms. As far as we know, this is therefore the first study that isolate and describe socioeconomic determinants of ME/CFS and calculate how these factors relate to the risk of ME/CFS diagnosis by utilizing individual level registry data. This allows for objective operationalization of the ME/CFS population, and makes it possible to model SES affect the risk of ME/CFS diagnosis, relative to control groups.

DATA AND METHODS

We conduct a pooled cross-sectional analysis of registry data from all adult patients diagnosed with ME/CFS from 2016 to 2018 in Norway, coupled with socioeconomic data from statistics Norway from 2011 to 2018. We operationalize SES as household income and educational attainment fixed at the beginning of the study period. We compare the effects of SES on the risk of ME/CFS diagnosis to a population of chronically ill patients with hospital diagnoses that share clinical characteristics of ME/CFS and a healthy random sample of the Norwegian population. Our models are estimated by logistic regression analyses.

RESULTS

When comparing the risk of ME/CFS diagnosis with a population consisting of people with four specific chronic diseases, we find that high educational attainment is associated with a 19% increase (OR: 1.19) in the risk of ME/CFS and that high household income is associated with a 17% decrease (OR:0.83) in risk of ME/CFS. In our second model we compare with a healthy population sample, and found that low educational attainment is associated with 69% decrease (OR:0.31) in the risk of ME/CFS and that low household income is associated with a 53% increase (OR: 1.53).

CONCLUSION

We find statistically significant associations between SES and the risk of ME/CFS. However, our more detailed analyses shows that our findings vary according to which population we compare the ME/CFS patients with, and that the effect of SES is larger when comparing with a healthy population sample, as opposed to controls with selected hospital diagnoses.

Temporal and age-specific trends in incidence and 1-year case-fatality of hospitalized ischaemic stroke in Victoria, Australia

OBJECTIVES

Investigate temporal and age-specific trends in the incidence of ischaemic stroke and case-fatality risk in Victoria, Australia.

MATERIALS AND METHODS

Patients hospitalised with first ischaemic stroke between 2012 and 2018 were included. Trends in age-standardised incidence rates of ischaemic stroke were assessed using linear regression models. Cox proportional hazard regression models were used to examine the case-fatality risk.

RESULTS

Overall age-standardised incidence of ischaemic stroke was stable from 2012/13 to 2017/18 (87.6 to 84.8 events per 100,000 population; Annual percentage change [APC] -0.32; 95% Confidence interval [CI] -1.13 to 0.50). The incidence declined in females (APC -1.00; 95% CI -1.49 to -0.50), people aged 75-84 years (APC -1.60; 95% CI -2.83 to -0.36) and in metropolitan areas (APC -0.74; 95% CI -1.02 to -0.45). The risk of 1-year case-fatality (HR 0.85; 95% CI 0.78 to 0.93) significantly declined in 2016/17 compared to 2012/13.

CONCLUSIONS

Overall ischaemic stroke incidence remained stable while decreasing trends were observed in females, elderly and metropolitan areas. 1-year case-fatality declined from 2012 to 2017.

Oral anticoagulant decreases stroke recurrence in patients with atrial fibrillation detected after stroke

Background

Atrial fibrillation detected after stroke (AFDAS) has a lower risk of ischemic stroke recurrence than known atrial fibrillation (KAF). While the benefit of oral anticoagulants (OAC) for preventing ischemic stroke recurrence in KAF is well established, their role in patients with AFDAS is more controversial. This study aimed to evaluate the association between OAC use and the risk of recurrent ischemic stroke in patients with AFDAS in a real-world setting.

Methods

This nationwide retrospective cohort study was conducted using the Taiwan National Health Insurance Research Database. Patients hospitalized with a first-ever ischemic stroke and AFDAS confirmed within 30 days after hospitalization were assigned to OAC and non-OAC cohorts. Inverse probability of treatment weighting was applied to balance the baseline characteristics of the cohorts. The primary outcome was ischemic stroke recurrence. Secondary outcomes were intracranial hemorrhage (ICH), death, and the composite outcome of "ischemic stroke recurrence, ICH, or death." Multivariate Cox proportional hazard models were used to estimate adjusted hazard ratios (aHR) and 95% confidence intervals (CI).

Results

A total of 4,508 hospitalized patients with stroke and AFDAS were identified. Based on OAC use, 2,856 and 1,652 patients were assigned to the OAC and non-OAC groups, respectively. During the follow-up period (median duration, 2.76 years), the OAC cohort exhibited a lower risk of ischemic stroke recurrence (aHR, 0.84; 95% CI, 0.70-0.99), death (aHR, 0.65; 95% CI, 0.58-0.73), and composite outcome (aHR, 0.70; 95% CI, 0.63-0.78) than did the non-OAC cohort. The risk of ICH (aHR, 0.96; 95% CI, 0.62-1.50) was not significantly different between the two cohorts.

Conclusion

OAC use in patients with AFDAS was associated with reduced risk of ischemic stroke recurrence, without an increased risk of ICH. This supports current guidelines recommending OACs for secondary stroke prevention in patients with AF, regardless of the time of diagnosis.

A nationwide registry study on heart failure in Norway from 2008 to 2018: variations in lookback period affect incidence estimates

Background

The incidence of heart failure (HF) has declined in Europe during the past two decades. However, incidence estimates from registry-based studies may vary, partly because they depend on retrospective searches to exclude previous events. The aim of this study was to assess to what extent different lookback periods (LPs) affect temporal trends in incidence, and to identify the minimal acceptable LP. Further, we wanted to estimate temporal trends in incidence and prevalence of HF in a nationwide population, using the minimal acceptable LP.

Methods

We identified all in- and out-patient contacts for HF in Norway during 2008 to 2018 from the Norwegian Patient Registry. To calculate the influence of varying LP on incident cases, we defined 2018 with 10 years of LP as a reference and calculated the relative difference by using one through 9 years of lookback. Temporal trends in incidence rates were estimated with sensitivity analyses applying varying LPs and different case definitions. Standardised incidence rates and prevalence were calculated by applying direct age- and sex-standardization to the 2013 European Standard Population.

Results

The overestimation of incident cases declined with increasing number of years included in the LP. Compared to a 10-year LP, application of 4, 6, and 8 years resulted in an overestimation of incident cases by 13.5%, 6.2% and 2.3%, respectively. Temporal trends in incidence were affected by the number of years in the LP and whether the LP was fixed or varied. Including all available data mislead to conclusions of declining incidence rates over time due to increasing LPs.

Conclusions

When taking the number of years with available data and HF mortality and morbidity into consideration, we propose that 6 years of fixed lookback is sufficient for identification of incident HF cases. HF incidence rates and prevalence increased from 2014 to 2018.

Trial registration

Retrospectively registered.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-022-02522-y.

The impact of lookback windows on the prevalence and incidence of chronic diseases among people living with HIV: an exploration in administrative health data in Canada

BACKGROUND

We described the impact of different lengths of lookback window (LW), a retrospective time period to observe diagnoses in administrative data, on the prevalence and incidence of eight chronic diseases.

METHODS

Our study populations included people living with HIV (N = 5151) and 1:5 age-sex-matched HIV-negative individuals (N = 25,755) in British Columbia, Canada, with complete follow-up between 1996 and 2012. We measured period prevalence and incidence of diseases in 2012 using LWs ranging from 1 to 16 years. Cases were deemed prevalent if identified in 2012 or within a defined LW, and incident if newly identified in 2012 with no previous cases detected within a defined LW. Chronic disease cases were ascertained using published case-finding algorithms applied to population-based provincial administrative health datasets.

RESULTS

Overall, using cases identified by the full 16-year LW as the reference, LWs ≥8 years and ≥ 4 years reduced the proportion of misclassified prevalent and incidence cases of most diseases to < 20%, respectively. The impact of LWs varied across diseases and populations.

CONCLUSIONS

This study underscored the importance of carefully choosing LWs and demonstrated data-driven approaches that may inform these choices. To improve comparability of prevalence and incidence estimates across different settings, we recommend transparent reporting of the rationale and limitations of chosen LWs.

The effects of different lookback periods on the sociodemographic structure of the study population and on the estimation of incidence rates: analyses with German claims data

BACKGROUND

Defining incident cases has always been a challenging issue for researchers working with routine data. Lookback periods should enable researchers to identify and exclude recurrent cases and increase the accuracy of the incidence estimation. There are different recommendations for lookback periods depending on a disease entity of up to 10 years. Well-known drawbacks of the application of lookback periods are shorter remaining observation period in the dataset or smaller number of cases. The problem of selectivity of the remaining population after introducing lookback periods has not been considered in the literature until now.

METHODS

The analyses were performed with pseudonymized claims data of a German statutory health insurance fund with annual case numbers of about 2,1 million insured persons. Proportions of study population excluded due to the application of lookback periods are shown according to age, occupational qualification and income. Myocardial infarction and stroke were used to demonstrate changes in incidence rates after applying lookback periods of up to 5 years.

RESULTS

Younger individuals show substantial dropouts after the application of lookback periods. Furthermore, there are selectivities regarding occupational qualification and income, which cannot be handled by age standardization. Due to selective dropouts of younger individuals, crude incidence rates of myocardial infarction and stroke increase after applying lookback periods. Depending on the income group, age-standardized incidence rates changed differentially, leading to a decrease and possible underestimation of the social gradient after applying lookback periods.

CONCLUSIONS

Selectivity analyses regarding age and sociodemographic structure should be performed for the study population after applying lookback periods since the selectivity can affect the outcome especially in health care research. The selectivity effects might occur not only in claims data of one health insurance fund, but also in other longitudinal data with left- or right-censoring not covering the whole population. The effects may also apply to health care systems with a mix of public and private health insurance. A trade-off has to be considered between selectivity effects and eliminating recurrent events for more accuracy in the definition of incidence.

Impact of variable look-back periods on the incidence rates of chronic diseases using real world data

PURPOSE

Estimating disease incidence based on secondary data requires a look-back period to exclude patients with pre-existing disease from the incidence risk set. However, the optimal length of the look-back period and its impact on incidence rates are often unknown. We studied the impact of the length of the look-back period on incidence rates of 24 different chronic diseases.

METHODS

Everyone residing in Sweden between January 1, 2005 and December 31, 2013 were identified from national registries and followed up to 2 years (through December 31, 2015). Outcome events were identified from inpatient and outpatient hospital contacts and incidence rates were calculated per 100 000 person-years. The length of the look-back period was varied with 6-month increments, starting at 6 months. The maximum look-back period of 9 years was used as reference period.

RESULTS

There were 7 943 807 individuals with a look-back period of at least 9 years (mean age 46.5 years) and a mean follow-up time of 1.97 years. Incidence rates were higher across all diseases when restricting the look-back to 1 year compared to 9 years, with a magnitude of overestimation of the incidence rates between 13% (temporal arteritis) and 174% (type 1 diabetes). However, for most diseases the effect of extending the look-back period beyond 3-5 years appeared comparably small.

CONCLUSIONS

This study illustrates how short look-back periods cause overestimation of the incidence rates of chronic diseases, suggesting that sensitivity analyses with respect to look-back period are considered, particularly using data sources with limited information on past medical history.

Incidence of Hospitalization for Stroke in Queensland, Australia: Younger Adults at Risk

BACKGROUND

Trends in the incidence of stroke are important for health care planning. Information is particularly scarce in Australia, due to the paucity of studies with access to recent, large-scale, longitudinal datasets. In this paper we investigated the incidence of hospitalization for stroke by sex, age, and subtype in the whole State of Queensland (Australia).

METHODS

We obtained data of all hospital admissions for stroke in Queensland from 2002 to 2015. Age standardized hospitalization rates for first-ever stroke were calculated along with WHO adjusted rates. Poisson regression analyses were conducted to investigate the influence of time and gender on the incidence of subtypes of stroke as well as the total incidence.

RESULTS

Admissions for first-ever stroke were 57,597. Crude hospitalization incidence rose from 87 (95% confidence interval [CI], 84-90) to 108 (95%CI, 105-111) for both sexes. The highest increase was in the age bands 40-49 from 33 (95%CI, 26-40) to 54 (95%CI, 46-62; +62%) and 50-59 from 82 (95%CI, 70-93) to 127 (95%CI, 114-140; +56%) in men and in 60-69 from 118 (95%CI, 100-136) to 159 (95%CI, 143-175; +34%) in women. Ischemic subtype rates appeared to increase more than haemorrhagic rates. Age range, sex, and year reliably predicted incidence rates.

CONCLUSIONS

If these trends are maintained, the data predict further increases, especially in males aged 40-59. With people apparently stroking earlier and mortality rates dropping, the healthcare system in Australia is faced with a rapidly increasing care burden.

Measuring stroke outcomes for 74 501 patients using linked administrative data: System-wide estimates and validation of 'home-time' as a surrogate measure of functional status

AIMS

Administrative data offer cost-effective, whole-of-population stroke surveillance yet the lack of validated measures of functional status is a shortcoming. The number of days spent living at home after stroke ('home-time') is a patient-centred outcome that can be objectively ascertained from administrative data. Population-based validation against both severity and outcome measures and for all subtypes is lacking. We aimed to report representative 'home-time' estimates and validate 'home-time' as a surrogate measure of functional status after stroke.

METHODS

Stroke hospitalisations from a state-wide census in New South Wales, Australia, from January 1, 2005 to March 31, 2014 were linked to prehospital data, poststroke admissions and deaths. We correlated 90-day 'home-time' with Glasgow Coma Scale (GCS) scores, measured upon a patient's initial contact with paramedics and Functional Independence Measure (FIM) scores, measured upon entry to rehabilitation after the acute hospital stroke admission. Negative binomial regressions identified predictors of 'home-time'.

RESULTS

Patients with stroke (N = 74 501) spent a median of 53 days living at home 90 days after the event. Median 'home-time' was 60 days after ischaemic stroke, 49 days after subarachnoid haemorrhage and 0 days after intracerebral haemorrhage. GCS and FIM scores significantly correlated with 'home-time' (P < .001). Women spent significantly less time at home compared with men after stroke, although being married increased 'home-time' after ischaemic stroke and subarachnoid haemorrhage.

CONCLUSIONS

These findings underscore the immediate and adverse impact of stroke. 'Home-time' measured using administrative data is a robust, replicable and valid patient-centred outcome enabling inexpensive population-based surveillance and system-wide quality assessment.

Population-based stroke surveillance using big data: state-wide epidemiological trends in admissions and mortality in New South Wales, Australia

OBJECTIVES

Epidemiological trends for major causes of death and disability, such as stroke, may be monitored using administrative data to guide public health initiatives and service delivery.

METHODS

We calculated admissions rates for ischaemic stroke, intracerebral haemorrhage and subarachnoid haemorrhage between 1 January 2005 and December 31^st, 2013 and rates of 30-day mortality and 365-day mortality in 30-day survivors to 31 December 2014 for patients aged 15 years or older from New South Wales, Australia. Annual Average Percentage Change in rates was estimated using negative binomial regression.

RESULTS

Of 81,703 eligible admissions, 64,047 (78.4%) were ischaemic strokes and 13,302 (16.3%) and 4,778 (5.8%) were intracerebral and subarachnoid haemorrhages, respectively. Intracerebral haemorrhage admissions significantly declined by an average of 2.2% annually (95% Confidence Interval = -3.5% to -0.9%) (p < 0.001). Thirty-day mortality rates significantly declined for ischaemic stroke (Average Percentage Change -2.9%, 95% Confidence Interval = -5.2% to -1.0%) (p = 0.004) and subarachnoid haemorrhage (Average Percentage Change = -2.6%, 95% Confidence Interval = -4.8% to -0.2%) (p = 0.04). Mortality at 365-days amongst 30-day survivors of ischaemic stroke and intracerebral haemorrhage was stable over time and increased in subarachnoid haemorrhage (Annual Percentage Change 6.2%, 95% Confidence Interval = -0.1% to 12.8%), although not significantly (p = 0.05).

DISCUSSION

Improved prevention may have underpinned declining intracerebral haemorrhage rates while survival gains suggest that innovations in care are being successfully translated. Mortality in patients surviving the acute period is unchanged and may be increasing for subarachnoid haemorrhage warranting investment in post-discharge care and secondary prevention.

The effect of the look-back period for estimating incidence using administrative data

Background

The look-back period is needed to define baseline population for estimating incidence. However, short look-back period is known to overestimate incidence of diseases misclassifying prevalent cases to incident cases. The purpose of this study is to evaluate the impact of the various length of look-back period on the observed incidences of uterine leiomyoma, endometriosis and adenomyosis, and to estimate true incidences considering the misclassification errors in the longitudinal administrative data in Korea.

Methods

A total of 319,608 women between 15 to 54 years of age in 2002 were selected from Korea National Health Insurance Services (KNHIS) cohort database. In order to minimize misclassification bias incurred when applying various length of look-back period, we used 11 years of claim data to estimate the incidence by equally setting the look-back period to 11 years for each year using prediction model. The association between the year of diagnosis and the number of prevalent cases with the misclassification rates by each look-back period was investigated. Based on the findings, prediction models on the proportion of misclassified incident cases were developed using multiple linear regression.

Results

The proportion of misclassified incident cases of uterine leiomyoma, endometriosis and adenomyosis were 32.8, 10.4 and 13.6% respectively for the one-year look-back period in 2003. These numbers decreased to 6.3% in uterine leiomyoma and − 0.8% in both endometriosis and adenomyosis using all available look-back periods (11 years) in 2013.

Conclusion

This study demonstrates approaches for estimating incidences considering the different proportion of misclassified cases for various length of look-back period. Although the prediction model used for estimation showed strong R-squared values, follow-up studies are required for validation of the study results.

Potentials, challenges, and limitations of the analysis of administrative data on vascular limb amputations in health care

Although more and more data on lower limb amputations are becoming available by leveraging the widening access to health care administrative databases, the applicability of these data for public health decisions is still limited. Problems can be traced back to methodological issues, how data are generated and to conceptual issues, namely, how data are interpreted in a multidimensional environment. The present review summarised all of the steps from converting the claims data of administrative databases into the analytical data and reviewed the wide array of sources of potential biases in the analysis of such data. The origins of uncertainty of administrative data analysis include uncontrolled confounding due to a lack of clinical data, the left- and right-censored nature of data collection, the non-standardized diagnosis/procedure-based data extraction methods (i.e., numerator/denominator problems) and additional methodological problems associated with temporal and spatial analyses. The existence of these methodological challenges in the administrative data-based analysis should not deter the analysts from using these data as a powerful tool in the armamentarium of clinical research. However, it must be done with caution and a thorough understanding and respect of the methodological limitations. In addition to this requirement, there is a profound need for pursuing further research on methodology and widening the search for other indicators (structural, process or outcome) that allow a deeper insight how the quality of vascular care may be assessed. Effective research using administrative data is based on strong collaboration in three domains, namely expertise in claims data handling and processing, the clinical field, and statistical analysis. The final interpretations of results and the countermeasures on the level of vascular care ought to be grounded on the integrity of research, open discussions and institutionalized mechanisms of science arbitration and honest brokering.

The impact of disease severity adjustment on hospital standardised mortality ratios: Results from a service-wide analysis of ischaemic stroke admissions using linked pre-hospital, admissions and mortality data

Background

Administrative data are used to examine variation in thirty-day mortality across health services in several jurisdictions. Hospital performance measurement may be error-prone as information about disease severity is not typically available in routinely collected data to incorporate into case-mix adjusted analyses. Using ischaemic stroke as a case study, we tested the extent to which accounting for disease severity impacts on hospital performance assessment.

Methods

We linked all recorded ischaemic stroke admissions between July, 2011 and June, 2014 to death registrations and a measure of stroke severity obtained at first point of patient contact with health services, across New South Wales, Australia’s largest health service jurisdiction. Thirty-day hospital standardised mortality ratios were adjusted for either comorbidities, as is typically done, or for both comorbidities and stroke severity. The impact of stroke severity adjustment on mortality ratios was determined using 95% and 99% control limits applied to funnel plots and by calculating the change in rank order of hospital risk adjusted mortality rates.

Results

The performance of the stroke severity adjusted model was superior to incorporating comorbidity burden alone (c-statistic = 0.82 versus 0.75; N = 17,700 patients, 176 hospitals). Concordance in outlier classification was 89% and 97% when applying 95% or 99% control limits to funnel plots, respectively. The sensitivity rates of outlier detection using comorbidity adjustment compared with gold-standard severity and comorbidity adjustment was 74% and 83% with 95% and 99% control limits, respectively. Corresponding positive predictive values were 74% and 91%. Hospital rank order of risk adjusted mortality rates shifted between 0 to 22 places with severity adjustment (Median = 4.0, Inter-quartile Range = 2–7).

Conclusions

Rankings of mortality rates varied widely depending on whether stroke severity was taken into account. Funnel plots yielded largely concordant results irrespective of severity adjustment and may be sufficiently accurate as a screening tool for assessing hospital performance.

Racial differences in recurrent ischemic stroke risk and recurrent stroke case fatality

OBJECTIVE

To determine black-white differences in 1-year recurrent stroke and 30-day case fatality after a recurrent stroke in older US adults.

METHODS

We conducted a retrospective cohort study using a 5% random sample of Medicare beneficiaries with fee-for-service health insurance coverage who were hospitalized for ischemic stroke between 1999 and 2013. Hazard ratios for recurrent ischemic stroke and risk ratios for 30-day case fatality comparing blacks to whites were calculated with adjustment for demographics, risk factors, and competing risk of death when appropriate.

RESULTS

Among 128,789 Medicare beneficiaries having an ischemic stroke (mean age 80 years [SD 8 years], 60.4% male), 11.1% were black. The incidence rate of recurrent ischemic stroke per 1,000 person-years for whites and blacks was 108 (95% confidence interval [CI], 106-111) and 154 (95% CI 147-162) , respectively. The multivariable-adjusted hazard ratio for recurrent stroke among blacks compared with whites was 1.36 (95% CI 1.29-1.44). The case fatality after recurrent stroke for blacks and whites was 21% (95% CI 21%-22%) and 16% (95% CI 15%-18%), respectively. The multivariable-adjusted relative risk for mortality within 30 days of a recurrent stroke among blacks compared with whites was 0.82 (95% CI 0.73-0.93).

CONCLUSION

The risk of stroke recurrence among older Americans hospitalized for ischemic stroke is higher for blacks compared to whites, while 30-day case fatality after recurrent stroke remains lower for blacks.