Validity of ascertainment of co-morbid illness using administrative databases: a systematic review

Understanding the biases to sepsis surveillance and quality assurance caused by inaccurate coding in administrative health data

PURPOSE

Timely and accurate data on the epidemiology of sepsis are essential to inform policy decisions and research priorities. We aimed to investigate the validity of inpatient administrative health data (IAHD) for surveillance and quality assurance of sepsis care.

METHODS

We conducted a retrospective validation study in a disproportional stratified random sample of 10,334 inpatient cases of age ≥ 15 years treated in 2015-2017 in ten German hospitals. The accuracy of coding of sepsis and risk factors for mortality in IAHD was assessed compared to reference standard diagnoses obtained by a chart review. Hospital-level risk-adjusted mortality of sepsis as calculated from IAHD information was compared to mortality calculated from chart review information.

RESULTS

ICD-coding of sepsis in IAHD showed high positive predictive value (76.9-85.7% depending on sepsis definition), but low sensitivity (26.8-38%), which led to an underestimation of sepsis incidence (1.4% vs. 3.3% for severe sepsis-1). Not naming sepsis in the chart was strongly associated with under-coding of sepsis. The frequency of correctly naming sepsis and ICD-coding of sepsis varied strongly between hospitals (range of sensitivity of naming: 29-71.7%, of ICD-diagnosis: 10.7-58.5%). Risk-adjusted mortality of sepsis per hospital calculated from coding in IAHD showed no substantial correlation to reference standard risk-adjusted mortality (r = 0.09).

CONCLUSION

Due to the under-coding of sepsis in IAHD, previous epidemiological studies underestimated the burden of sepsis in Germany. There is a large variability between hospitals in accuracy of diagnosing and coding of sepsis. Therefore, IAHD alone is not suited to assess quality of sepsis care.

Characteristics and Disease Burden of Patients With Eosinophilic Granulomatosis With Polyangiitis Initiating Mepolizumab in the United States

BACKGROUND/OBJECTIVE

Although the high disease burden associated with eosinophilic granulomatosis with polyangiitis (EGPA) has been established, the disease burden in patients initiating mepolizumab in real-world practice is poorly understood. This study aimed to assess characteristics and burden of real-world patients with EGPA initiating mepolizumab.

METHODS

This was a database study (GSK study ID: 214156) of US patients (≥12 years old) with EGPA and ≥1 mepolizumab claim (index date) identified from the Merative MarketScan Commercial and Medicare Supplemental Databases (November 1, 2015, to March 31, 2020). Outcomes assessed in the 12-month baseline period before index (inclusive) included patient characteristics, treatment use, EGPA relapses, asthma exacerbations, health care resource utilization, and costs.

RESULTS

In the 103 patients included (mean age, 51.1 years; 63.1% female), the most common manifestations were asthma (89.3%), chronic sinusitis (57.3%), and allergic rhinitis (43.7%). In total, 91.3% of patients had ≥1 oral corticosteroid (OCS) claim (median dose, 7.4 mg/d prednisone-equivalent), 45.6% were chronic OCS users (≥10 mg/d during the 90 days preindex), 99.0% had ≥1 EGPA-related relapse, and 62.1% ≥1 asthma exacerbation. During the baseline period, 26.2% and 97.1% of patients had EGPA-related inpatient admissions and office visits, respectively. Median all-cause total health care costs per patient were $33,298, with total outpatient costs ($16,452) representing the largest driver.

CONCLUSIONS

Before initiating mepolizumab, a substantial real-world EGPA disease burden is evident for patients, with resulting impact on health care systems, and indicative of unmet medical needs. Mepolizumab treatment, with a demonstrated positive clinical benefit-risk profile may represent a useful treatment option for reducing EGPA disease burden.

Viewpoint: how to measure comorbidities in patients with rheumatoid arthritis - clinical and academic value

Given the high prevalence and the enormous impact on key outcomes, comorbidities are important to consider, especially in patients with RA. Comorbidity indices are tools to quantify the impact of the overall burden of coexisting diseases on a specific outcome of interest. Until now, no gold standard exists on how to measure comorbidities. A large variety of indices have been developed using different settings and therefore leading to conceptual differences. Choosing the right tool clearly depends on the intention (clinical or research purpose) and the specific research question. The current article will address the purpose and challenge of measuring comorbidities in RA patients.

Validity of chronic disease diagnoses in Icelandic healthcare registries

AIMS

To evaluate the validity of recorded chronic disease diagnoses in Icelandic healthcare registries.

METHODS

Eight different chronic diseases from multiple sub-specialties of medicine were validated with respect to accuracy, but not to timeliness. For each disease, 30 patients with a recorded diagnosis and 30 patients without the same diagnosis were randomly selected from >80,000 participants in the iStopMM trial, which includes 54% of the Icelandic population born before 1976. Each case was validated by chart review by physicians using predefined criteria.

RESULTS

The overall accuracy of the chronic disease diagnoses was 96% (95% CI 94-97%), ranging from 92 to 98% for individual diseases. After weighting for disease prevalence, the accuracy was estimated to be 98.5%. The overall positive predictive value (PPV) of chronic disease diagnosis was 93% (95% CI 89-96%) and the overall negative predictive value (NPV) was 99% (95% CI 96-100%). There were disease-specific differences in validity, most notably multiple sclerosis, where the PPV was 83%. Other disorders had PPVs between 93 and 97%. The NPV of most disorders was 100%, except for hypertension and heart failure, where it was 97 and 93%, respectively. Those who had the registered chronic disease had objective findings of disease in 96% of cases.

CONCLUSIONS

When determining the presence of chronic disease, diagnosis data from the Icelandic healthcare registries has a high PPV, NPV and accuracy. Furthermore, most diagnoses can be confirmed by objective findings such as imaging or blood testing. These findings can inform the interpretation of studies using diagnostic data from the Icelandic healthcare registries.

Verification of administrative data to measure palliative care at terminal hospital stays

BACKGROUND

Administrative data and clinician documentation have not been directly compared for reporting palliative care, despite concerns about under-reporting.

OBJECTIVE

The aim of this study was to verify the use of routinely collected administrative data for reporting in-hospital palliation and to examine factors associated with coded palliative care in hospital administrative data.

METHOD

Hospital administrative data and inpatient palliative care activity documented in medical records were compared for patients dying in hospital between 1 July 2017 and 31 December 2017. Coding of palliative care in administrative data is based on hospital care type coded as "palliative care" and/or assignment of the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM) palliative care diagnosis code Z51.5. Medical records were searched for specified keywords, which, read in context, indicated a palliative approach to care. The list of keywords (palliative, end of life, comfort care, cease observations, crisis medications, comfort medications, syringe driver, pain or symptom management, no cardiopulmonary resuscitation, advance medical plan/resuscitation plan, deteriorating, agitation, restless and delirium) was developed in consultation with seven local clinicians specialising in palliative care or geriatric medicine.

RESULTS

Of the 576 patients who died in hospital, 246 were coded as having received palliative care, either solely by the ICD-10-AM diagnosis code Z51.5 (42%) or in combination with a "palliative care" care type (58%). Just over one-third of dying patients had a palliative care specialist involved in their hospital care. Involvement of a palliative care specialist and a cancer diagnosis substantially increased the odds of a Z51.5 code (odds ratio = 11 and 4, respectively). The majority of patients with a "syringe driver" or identified as being at the "end of life" were assigned a Z51.5 code (73.5% and 70.5%, respectively), compared to 53.8% and 54.7%, respectively, for "palliative" or "comfort care." For each keyword indicating a palliative approach to care, the Z51.5 code was more likely to be assigned if the patient had specialist palliative care input or if they had cancer.

CONCLUSION

Our results suggest administrative data under-represented in-hospital palliative care, at least partly due to medical record documentation that failed to meet ICD-10-AM coding criteria. Collaboration between clinicians and coders can enhance the quality of records and, consequently, administrative data.

The German Quality Network Sepsis: Evaluation of a Quality Collaborative on Decreasing Sepsis-Related Mortality in a Controlled Interrupted Time Series Analysis

Background

Sepsis is one of the leading causes of preventable deaths in hospitals. This study presents the evaluation of a quality collaborative, which aimed to decrease sepsis-related hospital mortality.

Methods

The German Quality Network Sepsis (GQNS) offers quality reporting based on claims data, peer reviews, and support for establishing continuous quality management and staff education. This study evaluates the effects of participating in the GQNS during the intervention period (April 2016–June 2018) in comparison to a retrospective baseline (January 2014–March 2016). The primary outcome was all-cause risk-adjusted hospital mortality among cases with sepsis. Sepsis was identified by International Classification of Diseases (ICD) codes in claims data. A controlled time series analysis was conducted to analyze changes from the baseline to the intervention period comparing GQNS hospitals with the population of all German hospitals assessed via the national diagnosis-related groups (DRGs)-statistics. Tests were conducted using piecewise hierarchical models. Implementation processes and barriers were assessed by surveys of local leaders of quality improvement teams.

Results

Seventy-four hospitals participated, of which 17 were university hospitals and 18 were tertiary care facilities. Observed mortality was 43.5% during baseline period and 42.7% during intervention period. Interrupted time-series analyses did not show effects on course or level of risk-adjusted mortality of cases with sepsis compared to the national DRG-statistics after the beginning of the intervention period (p = 0.632 and p = 0.512, respectively). There was no significant mortality decrease in the subgroups of patients with septic shock or ventilation >24 h or predefined subgroups of hospitals. A standardized survey among 49 local quality improvement leaders in autumn of 2018 revealed that most hospitals did not succeed in implementing a continuous quality management program or relevant measures to improve early recognition and treatment of sepsis. Barriers perceived most commonly were lack of time (77.6%), staff shortage (59.2%), and lack of participation of relevant departments (38.8%).

Conclusion

As long as hospital-wide sepsis quality improvement efforts will not become a high priority for the hospital leadership by assuring adequate resources and involvement of all pertinent stakeholders, voluntary initiatives to improve the quality of sepsis care will remain prone to failure.

Charlson Comorbidity Index: A Critical Review of Clinimetric Properties

The present critical review was conducted to evaluate the clinimetric properties of the Charlson Comorbidity Index (CCI), an assessment tool designed specifically to predict long-term mortality, with regard to its reliability, concurrent validity, sensitivity, incremental and predictive validity. The original version of the CCI has been adapted for use with different sources of data, ICD-9 and ICD-10 codes. The inter-rater reliability of the CCI was found to be excellent, with extremely high agreement between self-report and medical charts. The CCI has also been shown either to have concurrent validity with a number of other prognostic scales or to result in concordant predictions. Importantly, the clinimetric sensitivity of the CCI has been demonstrated in a variety of medical conditions, with stepwise increases in the CCI associated with stepwise increases in mortality. The CCI is also characterized by the clinimetric property of incremental validity, whereby adding the CCI to other measures increases the overall predictive accuracy. It has been shown to predict long-term mortality in different clinical populations, including medical, surgical, intensive care unit (ICU), trauma, and cancer patients. It may also predict in-hospital mortality, although in some instances, such as ICU or trauma patients, the CCI did not perform as well as other instruments designed specifically for that purpose. The CCI thus appears to be clinically useful not only to provide a valid assessment of the patient's unique clinical situation, but also to demarcate major diagnostic and prognostic differences among subgroups of patients sharing the same medical diagnosis.

Exploring the application of the Charlson Comorbidity Index to assess the patient population seen in a Veterans Affairs chiropractic residency program

OBJECTIVE

Chiropractic trainees require exposure to a diverse patient base, including patients with multiple medical conditions. The Veterans Affairs (VA) Chiropractic Residency Program aims for its doctor of chiropractic (DC) residents to gain experience managing a range of multimorbid cases, yet to our knowledge there are no published data on the comorbidity characteristics of patients seen by VA DC residents. We tested 2 approaches to obtaining Charlson Comorbidity Index (CCI) scores and compared CCI scores of resident patients with those of staff DCs at 1 VA medical center.

METHODS

Two processes of data collection to calculate CCI scores were developed. Time differences and agreement between methods were assessed. Comparison of CCI distribution between resident DC and staff DCs was done using 100 Monte Carlo simulation iterations of Fisher's exact test.

RESULTS

Both methods were able to calculate CCI scores (n = 22). The automated method was faster than the manual (13 vs 78 seconds per patient). CCI scores agreement between methods was good (κ = 0.67). We failed to find a significant difference in the distribution of resident DC and staff DC patients (mean p = .377; 95% CI, .375-.379).

CONCLUSION

CCI scores of a VA chiropractic resident's patients are measurable with both manual and automated methods, although automated may be preferred for its time efficiency. At the facility studied, the resident and staff DCs did not see patients with significantly different distributions of CCI scores. Applying CCI may give better insight into the characteristics of DC trainee patient populations.

[Opportunities and limitations of risk adjustment of quality indicators based on inpatient administrative health data - a workshop report]

INTRODUCTION

The quality indicators of the Initiative Qualitätsmedizin e. V. (IQM) have been developed as triggers to examine treatment processes for opportunities for improvement. Published quality results have partly been used for external quality comparisons in the media. Therefore, member hospitals of IQM demanded to investigate if methods of risk adjustment should be applied in the calculation of the quality indicators. After a hearing of experts had been held, a task force was founded to conduct test calculations on risk adjustment methods.

METHODS

Specific risk adjustment models for mortality in myocardial infarction, heart failure, stroke, pneumonia, and colectomy in colorectal cancer were developed in the database of national German DRG data of the year 2016. These models were used to calculate standardized mortality ratios (SMR) per indicator in a sample of 172 member hospitals of IQM based on the data of the year 2018. Median SMR per indicator were compared to median SMR based on a standardization by age and gender, which is the standard procedure in IQM. Correlations between the different SMR were calculated. Quality of care was judged by two different approaches: a) a descriptive discrepancy of |0.1| from the SMR value of 1, and b) a significant discrepancy from 1 using the 95% confidence limits. The effect of using the specific risk adjustment in relation to the standard procedure was investigated for both approaches (a and b).

RESULTS

The specific risk adjustment methods showed an area under the curve between 0.72 and 0.84. The median differences between the SMR based on standardization by age and gender and the SMR based on specific risk adjustment were small (between 0 and 0.4); Spearman's correlations were between 0.90 and 0.99. Changes in the judgement of quality of care in comparison to the national average occurred in 3.9% (mortality from pneumonia) to 20.6% of the hospitals (mortality from heart failure) in descriptive comparisons. When the judgement was based on confidence limits changes were observed in 1.6% (mortality after colectomy) to 17.4% of the hospitals (mortality from heart failure).

DISCUSSION

Implementing specific risk adjustment models had only minor effects on the distribution of risk-adjusted mortality compared to the standard procedure, but the judgement of quality of care could change for a fifth of the hospitals in individual indicators. Concerning methodological and practical reasons, the task force recommends further development of risk adjustment methods for selected indicators. This should be accompanied by studies on the validity of inpatient administrative data for quality management as well as by efforts to improve the usefulness of these data for such purposes.

Comorbidity and severity-of-illness risk adjustment for hospital-onset Clostridioides difficile infection using data from the electronic medical record

OBJECTIVE

To determine whether electronically available comorbidities and laboratory values on admission are risk factors for hospital-onset Clostridioides difficile infection (HO-CDI) across multiple institutions and whether they could be used to improve risk adjustment.

PATIENTS

All patients at least 18 years of age admitted to 3 hospitals in Maryland between January 1, 2016, and January 1, 2018.

METHODS

Comorbid conditions were assigned using the Elixhauser comorbidity index. Multivariable log-binomial regression was conducted for each hospital using significant covariates (P < .10) in a bivariate analysis. Standardized infection ratios (SIRs) were computed using current Centers for Disease Control and Prevention (CDC) risk adjustment methodology and with the addition of Elixhauser score and individual comorbidities.

RESULTS

At hospital 1, 314 of 48,057 patient admissions (0.65%) had a HO-CDI; 41 of 8,791 patient admissions (0.47%) at community hospital 2 had a HO-CDI; and 75 of 29,211 patient admissions (0.26%) at community hospital 3 had a HO-CDI. In multivariable regression, Elixhauser score was a significant risk factor for HO-CDI at all hospitals when controlling for age, antibiotic use, and antacid use. Abnormal leukocyte level at hospital admission was a significant risk factor at hospital 1 and hospital 2. When Elixhauser score was included in the risk adjustment model, it was statistically significant (P < .01). Compared with the current CDC SIR methodology, the SIR of hospital 1 decreased by 2%, whereas the SIRs of hospitals 2 and 3 increased by 2% and 6%, respectively, but the rankings did not change.

CONCLUSIONS

Electronically available patient comorbidities are important risk factors for HO-CDI and may improve risk-adjustment methodology.

Routine Preoperative Nutritional Screening in All Primary Total Joint Arthroplasty Patients Has Little Utility

BACKGROUND

Nutritional optimization before total joint arthroplasty (TJA) may improve patient outcomes and decrease costs. However, the utility of serologic laboratory markers, including albumin, transferrin, and total lymphocyte count (TLC), as primary indicators of nutrition is unclear. We analyzed the prevalence of abnormal nutritional values before TJA and identified factors associated with them.

METHODS

We retrospectively reviewed 819 primary cases of TJA performed at 1 institution from January to December 2018. Patient demographic characteristics were assessed for associations with abnormal preoperative nutritional values (albumin <3.5 g/dL, transferrin <200 mg/dL, and TLC <1.5 cells/μL³). Associations of comorbidities, American Society of Anesthesiologists Physical Status classification, and age-adjusted Charlson Comorbidity Index (CCI) with abnormal values were assessed with logistic regression.

RESULTS

Values were abnormal for albumin in 21 cases (2.6%), transferrin in 26 cases (5.6%), and TLC in 185 cases (25%). Thirteen cases (1.7%) had abnormal values for 2 markers. Age was associated with abnormal albumin and TLC, and race with abnormal transferrin. Congestive heart failure, chronic kidney disease, pancreatic insufficiency, gastroesophageal reflux disease, osteoporosis, dementia, and CCI were associated with abnormal albumin; Parkinson disease and American Society of Anesthesiologists Physical Status with abnormal transferrin; and dementia, body mass index, cancer history, and CCI with abnormal TLC.

CONCLUSION

We report low prevalence of and a low concordance rate among abnormal nutritional values before primary TJA. Our results suggest that routine testing of all healthy patients is not warranted before TJA.

Modified-Chronic Disease Score (M-CDS): Predicting the individual risk of death using drug prescriptions

Background

Estimating the morbidity of a population is strategic for health systems to improve healthcare. In recent years administrative databases have been increasingly used to predict health outcomes. In 1992, Von Korff proposed a Chronic Disease Score (CDS) to predict 1-year mortality by only using drug prescription data. Because pharmacotherapy underwent many changes over the last 3 decades, the original version of the CDS has limitations. The aim of this paper is to report on the development of the modified version of the CDS.

Methods

The modified CDS (M-CDS) was developed using 33 variables (from drug prescriptions within two-year before 01/01/2018) to predict one-year mortality in Bologna residents aged ≥50 years. The population was split into training and testing sets for internal validation. Score weights were estimated in the training set using Cox regression model with LASSO procedure for variables selection. The external validation was carried out on the Imola population. The predictive ability of M-CDS was assessed using ROC analysis and compared with that of the Charlson Comorbidity Index (CCI), that is based on hospital data only, and of the Multisource Comorbidity Score (MCS), which uses hospital and pharmaceutical data.

Results

The predictive ability of M-CDS was similar in the training and testing sets (AUC 95% CI: training [0.760–0.770] vs. testing [0.750–0.772]) and in the external population (Imola AUC 95% CI [0.756–0.781]). M-CDS was significantly better than CCI (M-CDS AUC = 0.761, 95% CI [0.750–0.772] vs. CCI-AUC = 0.696, 95% CI [0.681–0.711]). No significant difference was found between M-CDS and MCS (MCS AUC = 0.762, 95% CI [0.749–0.775]).

Conclusions

M-CDS, using only drug prescriptions, has a better performance than the CCI score in predicting 1-year mortality, and is not inferior to the multisource comorbidity score. M-CDS can be used for population risk stratification, for risk-adjustment in association studies and to predict the individual risk of death.

(Dis)concordance of comorbidity data and cancer status across administrative datasets, medical charts, and self-reports

BACKGROUND

Benchmarking outcomes across settings commonly requires risk-adjustment for co-morbidities that must be derived from extant sources that were designed for other purposes. A question arises as to the extent to which differing available sources for health data will be concordant when inferring the type and severity of co-morbidities, how close are these to the "truth". We studied the level of concordance for same-patient comorbidity data extracted from administrative data (coded from International Classification of Diseases, Australian modification,10th edition [ICD-10 AM]), from the medical chart audit, and data self-reported by men with prostate cancer who had undergone a radical prostatectomy.

METHODS

We included six hospitals (5 public and 1 private) contributing to the Prostate Cancer Outcomes Registry-Victoria (PCOR-Vic) in the study. Eligible patients from the PCOR-Vic underwent a radical prostatectomy between January 2017 and April 2018.Health Information Manager's in each hospital, provided each patient's associated administrative ICD-10 AM comorbidity codes. Medical charts were reviewed to extract comorbidity data. The self-reported comorbidity questionnaire (SCQ) was distributed through PCOR-Vic to eligible men.

RESULTS

The percentage agreement between the administrative data, medical charts and self-reports ranged from 92 to 99% in the 122 patients from the 217 eligible participants who responded to the questionnaire. The presence of comorbidities showed a poor level of agreement between data sources.

CONCLUSION

Relying on a single data source to generate comorbidity indices for risk-modelling purposes may fail to capture the reality of a patient's disease profile. There does not appear to be a 'gold-standard' data source for the collection of data on comorbidities.

Specialist Palliative Care Activity at an Acute Care Tertiary Hospital and Its Representation in Administrative Data

Objective:

To quantify and examine specialist palliative care (SPC) in-hospital activity and compare it to routinely collected administrative data on palliative care (PC).

Methods:

All patients discharged from a large acute care tertiary hospital in New South Wales, Australia, between July 1 and December 31, 2017, were identified from the hospital’s data warehouse. Administrative data were supplemented with information from the electronic medical record for hospital stays which were assigned the PC additional diagnosis code (Z51.5); had a “palliative care” care type; or included SPC consultation.

Results:

Of 34 653 hospital stays, 524 were coded as receiving PC—based on care type (43%) and/or diagnosis code Z51.5 (100%). Specialist palliative care provided 1717 consultations over 507 hospital stays. Patients had 2 (median; interquartile range: 1-4) consultations during an average stay of 15.3 days (SD 15.78; median 10); the first occurred 7.0 days (SD 12.13; median 3) after admission. Of patient stays with an SPC consultation, 70% were assigned the PC Z51.5 code; 60% were referred for symptom management; 68% had cancer. One hundred forty-one patients were under a palliative specialist—either from initial hospital admission (49.6%) or later in their stay.

Conclusions:

Palliative care specialists provide expert input into patient management, benefitting patients and other clinicians. Administrative data inadequately capture their involvement in patient care, especially consultations, and are therefore inappropriate for reporting SPC activity. Exclusion of information related to SPC activity results in an incomplete and distorted representation of PC services and fails to acknowledge the valuable contribution made by SPC.

How effective are population health surveys for estimating prevalence of chronic conditions compared to anonymised clinical data?

INTRODUCTION

Population health surveys are used to record person-reported outcome measures for chronic health conditions and provide a useful source of data when evaluating potential disease burdens. The reliability of survey-based prevalence estimates for chronic diseases is unclear nonetheless. This study applied methodological triangulation via a data linkage method to validate prevalence of selected chronic conditions (angina, myocardial infarction, heart failure, and asthma).

METHODS

Linked healthcare records were used for a combined cohort of 11,323 adults from the 2013 and 2014 sweeps of the Welsh Health Survey (WHS). The approach utilised consented survey data linked to primary and secondary care electronic health record (EHR) data back to 2002 within the Secure Anonymised Information Linkage (SAIL) Databank.

RESULTS

This descriptive study demonstrates validation of survey and clinical data using data linkage for selected chronic cardiovascular conditions and asthma with varied success. The results indicate that identifying cases for separate cardiovascular conditions was limited without specific medication codes for each condition, but more straightforward for asthma, where there was an extensive list of medications available. For asthma there was better agreement between prevalence estimates based on survey and clinical data as a result.

CONCLUSION

Whilst the results provide external validity for the WHS as an instrument for estimating the burden of chronic disease, they also indicate that a data linkage appproach can be used to produce comparable prevalence estimates using clinical data if a defined condition-specific set of clinical codes are available.

Reliability of comorbidity scores derived from administrative data in the tertiary hospital intensive care setting: a cross-sectional study

Background

Hospital reporting systems commonly use administrative data to calculate comorbidity scores in order to provide risk-adjustment to outcome indicators.

Objective

We aimed to elucidate the level of agreement between administrative coding data and medical chart review for extraction of comorbidities included in the Charlson Comorbidity Index (CCI) and Elixhauser Index (EI) for patients admitted to the intensive care unit of a university-affiliated hospital.

Method

We conducted an examination of a random cross-section of 100 patient episodes over 12 months (July 2012 to June 2013) for the 19 CCI and 30 EI comorbidities reported in administrative data and the manual medical record system. CCI and EI comorbidities were collected in order to ascertain the difference in mean indices, detect any systematic bias, and ascertain inter-rater agreement.

Results

We found reasonable inter-rater agreement (kappa (κ) coefficient ≥0.4) for cardiorespiratory and oncological comorbidities, but little agreement (κ<0.4) for other comorbidities. Comorbidity indices derived from administrative data were significantly lower than from chart review: −0.81 (95% CI − 1.29 to − 0.33; p=0.001) for CCI, and −2.57 (95% CI −4.46 to −0.68; p=0.008) for EI.

Conclusion

While cardiorespiratory and oncological comorbidities were reliably coded in administrative data, most other comorbidities were under-reported and an unreliable source for estimation of CCI or EI in intensive care patients. Further examination of a large multicentre population is required to confirm our findings.

Prevalence and impact of non-cardiovascular comorbidities among older adults hospitalized for non-ST segment elevation acute coronary syndrome

Background

There is a paucity of information on the prognostic importance of non-cardiovascular comorbidities (NCCs) among patients with non-ST-elevation acute coronary syndrome (NSTE-ACS). This study examined the prevalence and impact of NCCs on the length of stay (LOS) and mortality among older adults hospitalized for NSTE-ACS.

Methods

Among 1,488 older adults (mean age 79.4±8.4 years; 62.0% male) hospitalized for NSTE-ACS at a tertiary hospital in Melbourne, Australia, during 2013-2015, we collected data on comorbidities, LOS, and discharge outcomes. Thirteen NCCs were studied. Negative binomial and Cox proportional regression models were applied to examine the association between NCCs and LOS and in-hospital death, respectively.

Results

Approximately 53% of the patients had ≥1 NCCs. Diabetes and renal disease as well as anemia and renal disease co-existed more frequently than expected. Compared to having no NCCs, having one NCC was not associated with a significant increase in the likelihood of longer LOS [incidence rate ratio (IRR) 1.07; 95% CI: 0.99-1.15; P=0.085] or in-hospital death [hazard ratio (HR) 1.11; 95% CI: 0.65-1.90; P=0.707]. However, having ≥2 NCCs was associated with 22% and 79% increased likelihood of longer LOS (IRR 1.22, 95% CI: 1.11-1.33; P<0.001) and in-hospital death (HR 1.79, 95% CI: 1.06-3.03; P=0.029), respectively, compared to not having any NCC. Certain NCC dyads [e.g., chronic pulmonary disease (CPD) + renal disease] exhibited multiplicative effect such that their impact on patients' LOS or survival exceeded the sum of the individual effects of the component NCCs.

Conclusions

Over half of older patients hospitalized with NSTE-ACS had NCCs. A higher burden of NCCs correlated with increased LOS and lower survival. Contemporary ACS management guidelines need to recognize and incorporate protocols for the treatment of individuals with multiple chronic conditions to reduce the occurrence of adverse outcomes.

Electronically Available Comorbidities Should Be Used in Surgical Site Infection Risk Adjustment

Background

Healthcare-associated infections such as surgical site infections (SSIs) are used by the Centers for Medicare and Medicaid Services (CMS) as pay-for-performance metrics. Risk adjustment allows a fairer comparison of SSI rates across hospitals. Until 2016, Centers for Disease Control and Prevention (CDC) risk adjustment models for pay-for-performance SSI did not adjust for patient comorbidities. New 2016 CDC models only adjust for body mass index and diabetes.

Methods

We performed a multicenter retrospective cohort study of patients undergoing surgical procedures at 28 US hospitals. Demographic data and International Classification of Diseases, Ninth Revision codes were obtained on patients undergoing colectomy, hysterectomy, and knee and hip replacement procedures. Complex SSIs were identified by infection preventionists at each hospital using CDC criteria. Model performance was evaluated using measures of discrimination and calibration. Hospitals were ranked by SSI proportion and risk-adjusted standardized infection ratios (SIR) to assess the impact of comorbidity adjustment on public reporting.

Results

Of 45394 patients at 28 hospitals, 573 (1.3%) developed a complex SSI. A model containing procedure type, age, race, smoking, diabetes, liver disease, obesity, renal failure, and malnutrition showed good discrimination (C-statistic, 0.73) and calibration. When comparing hospital rankings by crude proportion to risk-adjusted ranks, 24 of 28 (86%) hospitals changed ranks, 16 (57%) changed by ≥2 ranks, and 4 (14%) changed by >10 ranks.

Conclusions

We developed a well-performing risk adjustment model for SSI using electronically available comorbidities. Comorbidity-based risk adjustment should be strongly considered by the CDC and CMS to adequately compare SSI rates across hospitals.

Tools for the Assessment of Comorbidity Burden in Rheumatoid Arthritis

Introduction

Comorbidities influence the prognosis, clinical outcomes, disease activity, and treatment response in rheumatoid arthritis (RA). RA patients have a high-comorbidity burden necessitating their study. Comorbidity indices are used to measure comorbidities and to study their impacts on different outcomes. A large number of such indices are used in clinical research. Some indices have been specifically developed in RA patients.

Aim

This review aims to provide an overview of generic and specific comorbidity indices commonly used in RA research.

Methods

We performed a critical literature review of comorbidity indices in RA using the PubMed database.

Results/discussion

This non-systematic literature review provides an overview of generic and specific comorbidity indices commonly used in RA studies. Some of the older but commonly used comorbidity indices like the Charlson comorbidity index and the Elixhauser comorbidity measure were primarily developed to estimate mortality risk from comorbid diseases. They were not specifically developed for RA patients but have been widely used in rheumatology comorbidity measurement. Of the many comorbidity indices available, only the rheumatic disease comorbidity index (RDCI) and the multimorbidity index have been specifically developed in RA patients. The functional comorbidity index was developed to look at functional disability and has been used in RA patients considering that morbidity is more important than mortality in such patients. While there is limited data comparing these indices, available evidence seems to favor the use of RDCI as it predicts mortality, hospitalization, disability, and healthcare utilization. The choice of the index, however, depends on several factors such as the population under study, outcome of interest, and sources of data. More research is needed to study the RA-specific comorbidity measures to make evidence-based recommendations for the choice of a comorbidity measure.

Electronically Available Comorbid Conditions for Risk Prediction of Healthcare-Associated Clostridium difficile Infection

OBJECTIVE To analyze whether electronically available comorbid conditions are risk factors for Centers for Disease Control and Prevention (CDC)-defined, hospital-onset Clostridium difficile infection (CDI) after controlling for antibiotic and gastric acid suppression therapy use. PATIENTS Patients aged ≥18 years admitted to the University of Maryland Medical Center between November 7, 2015, and May 31, 2017. METHODS Comorbid conditions were assessed using the Elixhauser comorbidity index. The Elixhauser comorbidity index and the comorbid condition components were calculated using the International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM) codes extracted from electronic medical records. Bivariate associations between CDI and potential covariates for multivariable regression, including antibiotic use, gastric acid suppression therapy use, as well as comorbid conditions, were estimated using log binomial multivariable regression. RESULTS After controlling for antibiotic use, age, proton-pump inhibitor use, and histamine-blocker use, the Elixhauser comorbidity index was a significant risk factor for predicting CDI. There was an increased risk of 1.26 (95% CI, 1.19-1.32) of having CDI for each additional Elixhauser point added to the total Elixhauser score. CONCLUSIONS An increase in Elixhauser score is associated with CDI. Our study and other studies have shown that comorbid conditions are important risk factors for CDI. Electronically available comorbid conditions and scores like the Elixhauser index should be considered for risk-adjustment of CDC CDI rates. Infect Control Hosp Epidemiol 2018;39:297-301.

The Charlson Comorbidity Index in Registry-based Research

BACKGROUND

Comorbidities may have an important impact on survival, and comorbidity scores are often implemented in studies assessing prognosis. The Charlson Comorbidity index is most widely used, yet several adaptations have been published, all using slightly different conversions of the International Classification of Diseases (ICD) coding.

OBJECTIVE

To evaluate which coding should be used to assess and quantify comorbidity for the Charlson Comorbidity Index for registry-based research, in particular if older ICD versions will be used.

METHODS

A systematic literature search was used to identify adaptations and modifications of the ICD-coding of the Charlson Comorbidity Index for general purpose in adults, published in English. Back-translation to ICD version 8 and version 9 was conducted by means of the ICD-code converter of Statistics Sweden.

RESULTS

In total, 16 studies were identified reporting ICD-adaptations of the Charlson Comorbidity Index. The Royal College of Surgeons in the United Kingdom combined 5 versions into an adapted and updated version which appeared appropriate for research purposes. Their ICD-10 codes were back-translated into ICD-9 and ICD-8 according to their proposed adaptations, and verified with previous versions of the Charlson Comorbidity Index.

CONCLUSION

Many versions of the Charlson Comorbidity Index are used in parallel, so clear reporting of the version, exact ICD- coding and weighting is necessary to obtain transparency and reproducibility in research. Yet, the version of the Royal College of Surgeons is up-to-date and easy-to-use, and therefore an acceptable co-morbidity score to be used in registry-based research especially for surgical patients.

Developing and validating a novel multisource comorbidity score from administrative data: a large population-based cohort study from Italy

OBJECTIVE

To develop and validate a novel comorbidity score (multisource comorbidity score (MCS)) predictive of mortality, hospital admissions and healthcare costs using multiple source information from the administrative Italian National Health System (NHS) databases.

METHODS

An index of 34 variables (measured from inpatient diagnoses and outpatient drug prescriptions within 2 years before baseline) independently predicting 1-year mortality in a sample of 500 000 individuals aged 50 years or older randomly selected from the NHS beneficiaries of the Italian region of Lombardy (training set) was developed. The corresponding weights were assigned from the regression coefficients of a Weibull survival model. MCS performance was evaluated by using an internal (ie, another sample of 500 000 NHS beneficiaries from Lombardy) and three external (each consisting of 500 000 NHS beneficiaries from Emilia-Romagna, Lazio and Sicily) validation sets. Discriminant power and net reclassification improvement were used to compare MCS performance with that of other comorbidity scores. MCS ability to predict secondary health outcomes (ie, hospital admissions and costs) was also investigated.

RESULTS

Primary and secondary outcomes progressively increased with increasing MCS value. MCS improved the net 1-year mortality reclassification from 27% (with respect to the Chronic Disease Score) to 69% (with respect to the Elixhauser Index). MCS discrimination performance was similar in the four regions of Italy we tested, the area under the receiver operating characteristic curves (95% CI) being 0.78 (0.77 to 0.79) in Lombardy, 0.78 (0.77 to 0.79) in Emilia-Romagna, 0.77 (0.76 to 0.78) in Lazio and 0.78 (0.77 to 0.79) in Sicily.

CONCLUSION

MCS seems better than conventional scores for predicting health outcomes, at least in the general population from Italy. This may offer an improved tool for risk adjustment, policy planning and identifying patients in need of a focused treatment approach in the everyday medical practice.

Swine exposure and methicillin-resistant Staphylococcus aureus infection among hospitalized patients with skin and soft tissue infections in Illinois: A ZIP code-level analysis

BACKGROUND

Methicillin-resistant Staphylococcus aureus (MRSA), a bacterial pathogen, is a predominant cause of skin and soft tissue infections (SSTI) in the United States. Swine-production facilities have been recognized as potential environmental reservoirs of MRSA. To better understand how swine production may contribute to MRSA infection, we evaluated the association between MRSA infection among SSTI inpatients and exposure measures derived from national swine inventory data.

METHODS

Based on adjusted odds ratios from logistic regression models, we evaluated the association between swine exposure metrics and MRSA infections among all Illinois inpatient hospitalizations for SSTI from January 2008 through July 2011. We also assessed if swine exposures had greater association with suspected community-onset MRSA (CO-MRSA) compared to suspected hospital-onset MRSA (HO-MRSA). Exposures were estimated using the Farm Location and Agricultural Production Simulator, generating the number of farms with greater than 1000 swine per residential ZIP code and the residential ZIP code-level swine density (swine/km²).

RESULTS

For every increase in 100 swine/km² within a residential ZIP code, the adjusted OR (aOR) for MRSA infection was 1.36 (95% CI: 1.28-1.45). For every additional large farm (i.e., >1000 swine) per ZIP code, the aOR for MRSA infection was 1.06 (95% CI: 1.04-1.07). The aOR for ZIP codes with any large farms compared to those with no large farms was 1.24 (95% CI: 1.19-1.29). We saw no evidence of an increased association for CO-MRSA compared to HO-MRSA with either continuous exposure metric (aORs=0.99), and observed inconsistent results across exposure categories.

CONCLUSIONS

These publicly-available, ecological exposure data demonstrated positive associations between swine exposure measures and individual-level MRSA infections among SSTI inpatients. Though it is difficult to draw definitive conclusions due to limitations of the data, these findings suggest that the risk of MRSA may increase based on indirect environmental exposure to swine production. Future research can address measurement error related to these data by improving exposure assessment precision, increased specification of MRSA strain, and better characterization of specific environmental exposure pathways.

A Systematic Review of Comorbidity Measurement Methods for Patients With Nontraumatic Brain Injury in Inpatient Rehabilitation Settings

This review summarizes comorbidity measurements used on patients with nontraumatic brain injury in inpatient rehabilitation and describes findings on measurement validation and comorbidity profiles. MEDLINE and MEDLINE In-Process, EMBASE, PsycINFO, the Cochrane Database of Systematic Reviews, Health, and Psychosocial Measurement Instruments were searched. Two reviewers screened results according to predefined inclusion and exclusion criteria. Population, statistical methods, comorbidity measurement, justification of its use, and results involving comorbidity were extracted using a standard table. Of 9476 articles retrieved, 16 were included. Comorbidity has been measured using various methods including the following: number and type within various classification systems, such as the International Disease Classification system, the Charlson comorbidity index, Centers for Medicare and Medicaid Services comorbidity tiers and patient comorbidity and complexity level values and subsets of diagnoses within nonadministrative data studies. No studies have assessed the predictive ability of the comorbidity measurements for inpatient rehabilitation outcomes in this population. Because comorbidities are common among the nontraumatic brain injury population, the predictive validity of comorbidity measurements should be assessed to determine the most appropriate measure to predict or risk adjust rehabilitation outcomes, which has implications for the development of clinical guidelines, and to inform health service research, planning, and delivery.

Multimorbidity in Australia: Comparing estimates derived using administrative data sources and survey data

Background

Estimating multimorbidity (presence of two or more chronic conditions) using administrative data is becoming increasingly common. We investigated (1) the concordance of identification of chronic conditions and multimorbidity using self-report survey and administrative datasets; (2) characteristics of people with multimorbidity ascertained using different data sources; and (3) whether the same individuals are classified as multimorbid using different data sources.

Methods

Baseline survey data for 90,352 participants of the 45 and Up Study—a cohort study of residents of New South Wales, Australia, aged 45 years and over—were linked to prior two-year pharmaceutical claims and hospital admission records. Concordance of eight self-report chronic conditions (reference) with claims and hospital data were examined using sensitivity (Sn), positive predictive value (PPV), and kappa (κ).The characteristics of people classified as multimorbid were compared using logistic regression modelling.

Results

Agreement was found to be highest for diabetes in both hospital and claims data (κ = 0.79, 0.78; Sn = 79%, 72%; PPV = 86%, 90%). The prevalence of multimorbidity was highest using self-report data (37.4%), followed by claims data (36.1%) and hospital data (19.3%). Combining all three datasets identified a total of 46 683 (52%) people with multimorbidity, with half of these identified using a single dataset only, and up to 20% identified on all three datasets. Characteristics of persons with and without multimorbidity were generally similar. However, the age gradient was more pronounced and people speaking a language other than English at home were more likely to be identified as multimorbid by administrative data.

Conclusions

Different individuals, with different combinations of conditions, are identified as multimorbid when different data sources are used. As such, caution should be applied when ascertaining morbidity from a single data source as the agreement between self-report and administrative data is generally poor. Future multimorbidity research exploring specific disease combinations and clusters of diseases that commonly co-occur, rather than a simple disease count, is likely to provide more useful insights into the complex care needs of individuals with multiple chronic conditions.

The Effect of Adding Comorbidities to Current Centers for Disease Control and Prevention Central-Line-Associated Bloodstream Infection Risk-Adjustment Methodology

BACKGROUND Risk adjustment is needed to fairly compare central-line-associated bloodstream infection (CLABSI) rates between hospitals. Until 2017, the Centers for Disease Control and Prevention (CDC) methodology adjusted CLABSI rates only by type of intensive care unit (ICU). The 2017 CDC models also adjust for hospital size and medical school affiliation. We hypothesized that risk adjustment would be improved by including patient demographics and comorbidities from electronically available hospital discharge codes. METHODS Using a cohort design across 22 hospitals, we analyzed data from ICU patients admitted between January 2012 and December 2013. Demographics and International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) discharge codes were obtained for each patient, and CLABSIs were identified by trained infection preventionists. Models adjusting only for ICU type and for ICU type plus patient case mix were built and compared using discrimination and standardized infection ratio (SIR). Hospitals were ranked by SIR for each model to examine and compare the changes in rank. RESULTS Overall, 85,849 ICU patients were analyzed and 162 (0.2%) developed CLABSI. The significant variables added to the ICU model were coagulopathy, paralysis, renal failure, malnutrition, and age. The C statistics were 0.55 (95% CI, 0.51-0.59) for the ICU-type model and 0.64 (95% CI, 0.60-0.69) for the ICU-type plus patient case-mix model. When the hospitals were ranked by adjusted SIRs, 10 hospitals (45%) changed rank when comorbidity was added to the ICU-type model. CONCLUSIONS Our risk-adjustment model for CLABSI using electronically available comorbidities demonstrated better discrimination than did the CDC model. The CDC should strongly consider comorbidity-based risk adjustment to more accurately compare CLABSI rates across hospitals. Infect Control Hosp Epidemiol 2017;38:1019-1024.

Association of β-Blockers With Functional Outcomes, Death, and Rehospitalization in Older Nursing Home Residents After Acute Myocardial Infarction

IMPORTANCE

Although β-blockers are a mainstay of treatment after acute myocardial infarction (AMI), these medications are commonly not prescribed for older nursing home residents after AMI, in part owing to concerns about potential functional harms and uncertainty of benefit.

OBJECTIVE

To study the association of β-blockers after AMI with functional decline, mortality, and rehospitalization among long-stay nursing home residents 65 years or older.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study of nursing home residents with AMI from May 1, 2007, to March 31, 2010, used national data from the Minimum Data Set, version 2.0, and Medicare Parts A and D. Individuals with β-blocker use before AMI were excluded. Propensity score-based methods were used to compare outcomes in people who did vs did not initiate β-blocker therapy after AMI hospitalization.

MAIN OUTCOMES AND MEASURES

Functional decline, death, and rehospitalization in the first 90 days after AMI. Functional status was measured using the Morris scale of independence in activities of daily living.

RESULTS

The initial cohort of 15 720 patients (11 140 women [70.9%] and 4580 men [29.1%]; mean [SD] age, 83 [8] years) included 8953 new β-blocker users and 6767 nonusers. The propensity-matched cohort included 5496 new users of β-blockers and an equal number of nonusers for a total cohort of 10 992 participants (7788 women [70.9%]; 3204 men [29.1%]; mean [SD] age, 84 [8] years). Users of β-blockers were more likely than nonusers to experience functional decline (odds ratio [OR], 1.14; 95% CI, 1.02-1.28), with a number needed to harm of 52 (95% CI, 32-141). Conversely, β-blocker users were less likely than nonusers to die (hazard ratio [HR], 0.74; 95% CI, 0.67-0.83) and had similar rates of rehospitalization (HR, 1.06; 95% CI, 0.98-1.14). Nursing home residents with moderate or severe cognitive impairment or severe functional dependency were particularly likely to experience functional decline from β-blockers (OR, 1.34; 95% CI, 1.11-1.61 and OR, 1.32; 95% CI, 1.10-1.59, respectively). In contrast, little evidence of functional decline due to β-blockers was found in participants with intact cognition or mild dementia (OR, 1.03; 95% CI, 0.89-1.20; P = .03 for effect modification) or in those in the best (OR, 0.99; 95% CI, 0.77-1.26) and intermediate (OR, 1.05; 95% CI, 0.86-1.27) tertiles of functional independence (P = .06 for effect modification). Mortality benefits of β-blockers were similar across all subgroups.

CONCLUSIONS AND RELEVANCE

Use of β-blockers after AMI is associated with functional decline in older nursing home residents with substantial cognitive or functional impairment, but not in those with relatively preserved mental and functional abilities. Use of β-blockers yielded a considerable mortality benefit in all groups.

Influence of Using Different Databases and 'Look Back' Intervals to Define Comorbidity Profiles for Patients with Newly Diagnosed Hypertension: Implications for Health Services Researchers

OBJECTIVE

To determine the data sources and 'look back' intervals to define comorbidities.

DATA SOURCES

Hospital discharge abstracts database (DAD), physician claims, population registry and death registry from April 1, 1994 to March 31, 2010 in Alberta, Canada.

STUDY DESIGN

Newly-diagnosed hypertension cases from 1997 to 2008 fiscal years were identified and followed up to 12 years. We defined comorbidities using data sources and duration of retrospective observation (6 months, 1 year, 2 years, and 3 years). The C-statistics for logistic regression and concordance index (CI) for Cox model of mortality and cardiovascular disease hospitalization were used to evaluate discrimination performance for each approach of defining comorbidities.

PRINCIPAL FINDINGS

The comorbidities prevalence became higher with a longer duration. Using DAD alone underestimated the prevalence by about 75%, compared to using both DAD and physician claims. The C-statistic and CI were highest when both DAD and physician claims were used, and model performance improved when observation duration increased from 6 months to one year or longer.

CONCLUSION

The comorbidities prevalence is greatly impacted by the data source and duration of retrospective observation. A combination of DAD and physicians claims with at least one year observation duration improves predictions for cardiovascular disease and one-year mortality outcome model performance.

Defining and validating comorbidities and procedures in ICD-10 health data in ST-elevation myocardial infarction patients

Administrative health databases are used in research to define comorbid conditions, diagnosis, and procedures. Our objectives were to validate a diagnosis of ST-elevation myocardial infarction (STEMI) and invasive cardiac procedure coding against a comprehensive registry of STEMI patients and determine an optimal algorithm for defining comorbidities using administrative hospitalization and ambulatory databases, but without using a physician claims database, which is unavailable for use in many jurisdictions.A registry of consecutive STEMI patients was used to define a reference cohort and linked to the hospitalization and ambulatory databases. Four administrative case definitions for defining comorbidities, as well as STEMI diagnosis and in-hospital procedures using the International Classification of Diseases, 10th Revision (ICD-10) and the Canadian Classification of Health Interventions (CCI) were evaluated. Metrics were used to evaluate algorithm performance and compare discriminative ability using the C statistic.The 3236 patients had median age of 60 years (interquartile range 52-71) and 75.7% were male. A diagnosis of STEMI was correctly identified in the administrative records for 3043 (94.0%) patients. In-hospital procedures (coronary artery bypass grafting, percutaneous coronary intervention, and angiogram) were well identified using administrative definitions (Kappa statistic 0.83-1.00). Validation of comorbidities varied by condition but an algorithm using 2 inpatient/ambulatory visits in the previous 2 years maximized PPV, ranging from 28.6% for previous heart failure to 95.7% for previous MI. The c statistic was similar for each of the methods, ranging from 0.76 to 0.80.ICD-10 and CCI codes can identify hospitalized STEMI patients with high sensitivity and accurately define in-hospital cardiac procedures. Comorbidities can be defined with high PPV using a definition of 2 inpatient/ambulatory visits in the previous 2 years.

Administrative data is as good as medical chart review for comorbidity ascertainment in patients with infections in Singapore

The Charlson comorbidity index (CCI) is widely used for control of confounding from comorbidities in epidemiological studies. International Classification of Diseases (ICD)-coded diagnoses from administrative hospital databases is potentially an efficient way of deriving CCI. However, no studies have evaluated its validity in infectious disease research. We aim to compare CCI derived from administrative data and medical record review in predicting mortality in patients with infections. We conducted a cross-sectional study on 199 inpatients. Correlation analyses were used to compare comorbidity scores from ICD-coded administrative databases and medical record review. Multivariable regression models were constructed and compared for discriminatory power for 30-day in-hospital mortality. Overall agreement was fair [weighted kappa 0·33, 95% confidence interval (CI) 0·23-0·43]. Kappa coefficient ranged from 0·17 (95% CI 0·01-0·36) for myocardial infarction to 0·85 (95% CI 0·59-1·00) for connective tissue disease. Administrative data-derived CCI was predictive of CCI ⩾5 from medical record review, controlling for age, gender, resident status, ward class, clinical speciality, illness severity, and infection source (C = 0·773). Using the multivariable model comprising age, gender, resident status, ward class, clinical speciality, illness severity, and infection source to predict 30-day in-hospital mortality, administrative data-derived CCI (C = 0·729) provided a similar C statistic as medical record review (C = 0·717, P = 0·8548). In conclusion, administrative data-derived CCI can be used for assessing comorbidities and confounding control in infectious disease research.

Optimization is required when using linked hospital and laboratory data to investigate respiratory infections

OBJECTIVE

Despite a recommendation for microbiological testing, only 45% of children hospitalized for respiratory infections in our previous data linkage study linked to a microbiological record. We conducted a chart review to validate linked microbiological data.

STUDY DESIGN AND SETTING

The chart review consisted of children aged <5 years admitted to seven selected hospitals for respiratory infections in Western Australia, 2000-2011. We calculated the proportion of admissions where testing was performed and any pathogens detected. We compared these proportions between the chart review and our previous data linkage study. Poisson regression was used to identify factors predicting the likelihood of microbiological tests in the chart review cohort.

RESULTS

From the chart review, 77% of 746 records had a microbiological test performed compared with 46% of 18,687 records from our previous data linkage study. Of those undergoing testing, 66% of the chart review and 64% of data linkage records had ≥1 respiratory pathogen(s) detected. In the chart review cohort, frequency of testing was highest in children admitted to metropolitan hospitals.

CONCLUSION

Validation studies are essential to ensure the quality of linked data. Our previous data linkage study failed to capture all relevant microbiological records. Findings will be used to optimize extraction protocols for future linkage studies.

Charlson index scores from administrative data and case-note review compared favourably in a renal disease cohort

BACKGROUND

The Charlson index is a widely used measure of comorbidity. The objective was to compare Charlson index scores calculated using administrative data to those calculated using case-note review (CNR) in relation to all-cause mortality and initiation of renal replacement therapy (RRT) in the Grampian Laboratory Outcomes Mortality and Morbidity Study (GLOMMS-1) chronic kidney disease cohort.

METHODS

Modified Charlson index scores were calculated using both data sources in the GLOMMS-1 cohort. Agreement between scores was assessed using the weighted Kappa. The association with outcomes was assessed using Poisson regression, and the performance of each was compared using net reclassification improvement.

RESULTS

Of 3382 individuals, median age 78.5 years, 56% female, there was moderate agreement between scores derived from the two data sources (weighted kappa 0.41). Both scores were associated with mortality independent of a number of confounding factors. Administrative data Charlson scores were more strongly associated with death than CNR scores using net reclassification improvement. Neither score was associated with commencing RRT.

CONCLUSION

Despite only moderate agreement, modified Charlson index scores from both data sources were associated with mortality. Neither was associated with commencing RRT. Administrative data compared favourably and may be superior to CNR when used in the Charlson index to predict mortality.

Development of a Novel Electronic Surveillance System for Monitoring of Bloodstream Infections

Background.

Electronic surveillance systems (ESSs) that utilize existing information in databases are more efficient than conventional infection surveillance methods.

Objective.

To develop an ESS for monitoring bloodstream infections (BSIs) and assess whether data obtained from the ESS were in agreement with data obtained by traditional manual medical-record review.

Methods.

An ESS was developed by linking data from regional laboratory and hospital administrative databases. Definitions for excluding BSI episodes representing contamination and duplicate episodes were developed and applied. Infections were classified as nosocomial infections, healthcare-associated community-onset infections, or community-acquired infections. For a random sample of episodes, data in the ESS were compared with data obtained by independent medical chart review.

Results.

From the records of the 306 patients whose infections were selected for comparative review, the ESS identified 323 episodes of BSI, of which 107 (33%) were classified as healthcare-associated community-onset infections, 108 (33%) were classified as community-acquired infections, 107 (33%) were classified as nosocomial infections, and 1 (0.3%) could not be classified. In comparison, 310 episodes were identified by use of medical chart review, of which 116 (37%) were classified as healthcare-associated community-onset infections, 95 (31%) as community-acquired infections, and 99 (32%) as nosocomial infections. For 302 episodes of BSI, there was concordance between the findings of the ESS and those of traditional manual chart review. Of the additional 21 discordant episodes that were identified by use of the ESS, 17 (81%) were classified as representing isolation of skin contaminants, by use of chart review. Of the additional 8 discordant episodes further identified by use of chart review, most were classified as repeat or polymicrobial episodes of disease. There was an overall 85% agreement between the findings of the ESS and those of chart review (K = 0.78; standard error, K = 0.04) for classification according to location of acquisition.

Conclusion.

Our novel ESS allows episodes of BSI to be identified and classified with a high degree of accuracy. This system requires validation in other cohorts and settings.

Is routine hospital episode data sufficient for identifying individuals with chronic kidney disease? A comparison study with laboratory data

Internationally, investment in the availability of routine health care data for improving health, health surveillance and health care is increasing. We assessed the validity of hospital episode data for identifying individuals with chronic kidney disease compared to biochemistry data in a large population-based cohort, the Grampian Laboratory Outcomes, Morbidity and Mortality Study-II (n = 70,435). Grampian Laboratory Outcomes, Morbidity and Mortality Study-II links hospital episode data to biochemistry data for all adults in a health region with impaired kidney function and random samples of individuals with normal and unmeasured kidney function in 2003. We compared identification of individuals with chronic kidney disease by hospital episode data (based on International Classification of Diseases-10 codes) to the reference standard of biochemistry data (at least two estimated glomerular filtration rates <60 mL/min/1.73 m(2) at least 90 days apart). Hospital episode data, compared to biochemistry data, identified a lower prevalence of chronic kidney disease and had low sensitivity (<10%) but high specificity (>97%). Using routine health care data from multiple sources offers the best opportunity to identify individuals with chronic kidney disease.

Can valid and practical risk-prediction or casemix adjustment models, including adjustment for comorbidity, be generated from English hospital administrative data (Hospital Episode Statistics)? A national observational study

Background

NHS hospitals collect a wealth of administrative data covering accident and emergency (A&E) department attendances, inpatient and day case activity, and outpatient appointments. Such data are increasingly being used to compare units and services, but adjusting for risk is difficult.

Objectives

To derive robust risk-adjustment models for various patient groups, including those admitted for heart failure (HF), acute myocardial infarction, colorectal and orthopaedic surgery, and outcomes adjusting for available patient factors such as comorbidity, using England’s Hospital Episode Statistics (HES) data. To assess if more sophisticated statistical methods based on machine learning such as artificial neural networks (ANNs) outperform traditional logistic regression (LR) for risk prediction. To update and assess for the NHS the Charlson index for comorbidity. To assess the usefulness of outpatient data for these models.

Main outcome measures

Mortality, readmission, return to theatre, outpatient non-attendance. For HF patients we considered various readmission measures such as diagnosis-specific and total within a year.

Methods

We systematically reviewed studies comparing two or more comorbidity indices. Logistic regression, ANNs, support vector machines and random forests were compared for mortality and readmission. Models were assessed using discrimination and calibration statistics. Competing risks proportional hazards regression and various count models were used for future admissions and bed-days.

Results

Our systematic review and empirical analysis suggested that for general purposes comorbidity is currently best described by the set of 30 Elixhauser comorbidities plus dementia. Model discrimination was often high for mortality and poor, or at best moderate, for other outcomes, for examplec = 0.62 for readmission andc = 0.73 for death following stroke. Calibration was often good for procedure groups but poorer for diagnosis groups, with overprediction of low risk a common cause. The machine learning methods we investigated offered little beyond LR for their greater complexity and implementation difficulties. For HF, some patient-level predictors differed by primary diagnosis of readmission but not by length of follow-up. Prior non-attendance at outpatient appointments was a useful, strong predictor of readmission. Hospital-level readmission rates for HF did not correlate with readmission rates for non-HF; hospital performance on national audit process measures largely correlated only with HF readmission rates.

Conclusions

Many practical risk-prediction or casemix adjustment models can be generated from HES data using LR, though an extra step is often required for accurate calibration. Including outpatient data in readmission models is useful. The three machine learning methods we assessed added little with these data. Readmission rates for HF patients should be divided by diagnosis on readmission when used for quality improvement.

Future work

As HES data continue to develop and improve in scope and accuracy, they can be used more, for instance A&E records. The return to theatre metric appears promising and could be extended to other index procedures and specialties. While our data did not warrant the testing of a larger number of machine learning methods, databases augmented with physiological and pathology information, for example, might benefit from methods such as boosted trees. Finally, one could apply the HF readmissions analysis to other chronic conditions.

Funding

The National Institute for Health Research Health Services and Delivery Research programme.

Variation in the recording of common health conditions in routine hospital data: study using linked survey and administrative data in New South Wales, Australia

OBJECTIVES

To investigate the nature and potential implications of under-reporting of morbidity information in administrative hospital data.

SETTING AND PARTICIPANTS

Retrospective analysis of linked self-report and administrative hospital data for 32,832 participants in the large-scale cohort study (45 and Up Study), who joined the study from 2006 to 2009 and who were admitted to 313 hospitals in New South Wales, Australia, for at least an overnight stay, up to a year prior to study entry.

OUTCOME MEASURES

Agreement between self-report and recording of six morbidities in administrative hospital data, and between-hospital variation and predictors of positive agreement between the two data sources.

RESULTS

Agreement between data sources was good for diabetes (κ=0.79); moderate for smoking (κ=0.59); fair for heart disease, stroke and hypertension (κ=0.40, κ=0.30 and κ =0.24, respectively); and poor for obesity (κ=0.09), indicating that a large number of individuals with self-reported morbidities did not have a corresponding diagnosis coded in their hospital records. Significant between-hospital variation was found (ranging from 8% of unexplained variation for diabetes to 22% for heart disease), with higher agreement in public and large hospitals, and hospitals with greater depth of coding.

CONCLUSIONS

The recording of six common health conditions in administrative hospital data is highly variable, and for some conditions, very poor. To support more valid performance comparisons, it is important to stratify or control for factors that predict the completeness of recording, including hospital depth of coding and hospital type (public/private), and to increase efforts to standardise recording across hospitals. Studies using these conditions for risk adjustment should also be cautious of their use in smaller hospitals.

Approaches to ascertaining comorbidity information: validation of routine hospital episode data with clinician-based case note review

Background

In clinical practice, research, and increasingly health surveillance, planning and costing, there is a need for high quality information to determine comorbidity information about patients. Electronic, routinely collected healthcare data is capturing increasing amounts of clinical information as part of routine care. The aim of this study was to assess the validity of routine hospital administrative data to determine comorbidity, as compared with clinician-based case note review, in a large cohort of patients with chronic kidney disease.

Methods

A validation study using record linkage. Routine hospital administrative data were compared with clinician-based case note review comorbidity data in a cohort of 3219 patients with chronic kidney disease. To assess agreement, we calculated prevalence, kappa statistic, sensitivity, specificity, positive predictive value and negative predictive value. Subgroup analyses were also performed.

Results

Median age at index date was 76.3 years, 44% were male, 67% had stage 3 chronic kidney disease and 31% had at least three comorbidities. For most comorbidities, we found a higher prevalence recorded from case notes compared with administrative data. The best agreement was found for cerebrovascular disease (κ = 0.80) ischaemic heart disease (κ = 0.63) and diabetes (κ = 0.65). Hypertension, peripheral vascular disease and dementia showed only fair agreement (κ = 0.28, 0.39, 0.38 respectively) and smoking status was found to be poorly recorded in administrative data. The patterns of prevalence across subgroups were as expected and for most comorbidities, agreement between case note and administrative data was similar. Agreement was less, however, in older ages and for those with three or more comorbidities for some conditions.

Conclusions

This study demonstrates that hospital administrative comorbidity data compared moderately well with case note review data for cerebrovascular disease, ischaemic heart disease and diabetes, however there was significant under-recording of some other comorbid conditions, and particularly common risk factors.

Systematic review of comorbidity indices for administrative data

BACKGROUND

Adjustment for comorbidities is common in performance benchmarking and risk prediction. Despite the exponential upsurge in the number of articles citing or comparing Charlson, Elixhauser, and their variants, no systematic review has been conducted on studies comparing comorbidity measures in use with administrative data. We present a systematic review of these multiple comparison studies and introduce a new meta-analytical approach to identify the best performing comorbidity measures/indices for short-term (inpatient + ≤ 30 d) and long-term (outpatient+>30 d) mortality.

METHODS

Articles up to March 18, 2011 were searched based on our predefined terms. The bibliography of the chosen articles and the relevant reviews were also searched and reviewed. Multiple comparisons between comorbidity measures/indices were split into all possible pairs. We used the hypergeometric test and confidence intervals for proportions to identify the comparators with significantly superior/inferior performance for short-term and long-term mortality. In addition, useful information such as the influence of lookback periods was extracted and reported.

RESULTS

Out of 1312 retrieved articles, 54 articles were eligible. The Deyo variant of Charlson was the most commonly referred comparator followed by the Elixhauser measure. Compared with baseline variables such as age and sex, comorbidity adjustment methods seem to better predict long-term than short-term mortality and Elixhauser seems to be the best predictor for this outcome. For short-term mortality, however, recalibration giving empirical weights seems more important than the choice of comorbidity measure.

CONCLUSIONS

The performance of a given comorbidity measure depends on the patient group and outcome. In general, the Elixhauser index seems the best so far, particularly for mortality beyond 30 days, although several newer, more inclusive measures are promising.

The predictive value of ICD-10 diagnostic coding used to assess Charlson comorbidity index conditions in the population-based Danish National Registry of Patients

BACKGROUND

The Charlson comorbidity index is often used to control for confounding in research based on medical databases. There are few studies of the accuracy of the codes obtained from these databases. We examined the positive predictive value (PPV) of the ICD-10 diagnostic coding in the Danish National Registry of Patients (NRP) for the 19 Charlson conditions.

METHODS

Among all hospitalizations in Northern Denmark between 1 January 1998 and 31 December 2007 with a first-listed diagnosis of a Charlson condition in the NRP, we selected 50 hospital contacts for each condition. We reviewed discharge summaries and medical records to verify the NRP diagnoses, and computed the PPV as the proportion of confirmed diagnoses.

RESULTS

A total of 950 records were reviewed. The overall PPV for the 19 Charlson conditions was 98.0% (95% CI; 96.9, 98.8). The PPVs ranged from 82.0% (95% CI; 68.6%, 91.4%) for diabetes with diabetic complications to 100% (one-sided 97.5% CI; 92.9%, 100%) for congestive heart failure, peripheral vascular disease, chronic pulmonary disease, mild and severe liver disease, hemiplegia, renal disease, leukaemia, lymphoma, metastatic tumour, and AIDS.

CONCLUSION

The PPV of NRP coding of the Charlson conditions was consistently high.

The predictive value of ICD-10 diagnostic coding used to assess Charlson comorbidity index conditions in the population-based Danish National Registry of Patients

BACKGROUND

The Charlson comorbidity index is often used to control for confounding in research based on medical databases. There are few studies of the accuracy of the codes obtained from these databases. We examined the positive predictive value (PPV) of the ICD-10 diagnostic coding in the Danish National Registry of Patients (NRP) for the 19 Charlson conditions.

METHODS

Among all hospitalizations in Northern Denmark between 1 January 1998 and 31 December 2007 with a first-listed diagnosis of a Charlson condition in the NRP, we selected 50 hospital contacts for each condition. We reviewed discharge summaries and medical records to verify the NRP diagnoses, and computed the PPV as the proportion of confirmed diagnoses.

RESULTS

A total of 950 records were reviewed. The overall PPV for the 19 Charlson conditions was 98.0% (95% CI; 96.9, 98.8). The PPVs ranged from 82.0% (95% CI; 68.6%, 91.4%) for diabetes with diabetic complications to 100% (one-sided 97.5% CI; 92.9%, 100%) for congestive heart failure, peripheral vascular disease, chronic pulmonary disease, mild and severe liver disease, hemiplegia, renal disease, leukaemia, lymphoma, metastatic tumour, and AIDS.

CONCLUSION

The PPV of NRP coding of the Charlson conditions was consistently high.