Administrative Data Research Articles

Patient-Physician Sex Discordance and "Before Medically Advised" Discharge from Hospital: A Population-Based Retrospective Cohort Study.

Patient-physician sex discordance (when patient sex does not match physician sex) has been associated with reduced clinical rapport and adverse outcomes including post-operative mortality and unplanned hospital readmission. It remains unknown whether patient-physician sex discordance is associated with "before medically advised" hospital discharge (BMA discharge; commonly known as discharge "against medical advice"). To evaluate whether patient-physician sex discordance is associated with BMA discharge. Retrospective cohort study using 15years (2002-2017) of linked population-based administrative health data for all non-elective, non-obstetrical acute care hospitalizations from British Columbia, Canada. All individuals with eligible hospitalizations during study interval. Exposure: patient-physician sex discordance. BMA discharge (primary), 30-day hospital readmission or death (secondary). We identified 1,926,118 eligible index hospitalizations, 2.6% of which ended in BMA discharge. Among male patients, sex discordance was associated with BMA discharge (crude rate, 4.0% vs 2.9%; adjusted odds ratio [aOR] 1.08; 95%CI 1.03-1.14; p = 0.003). Among female patients, sex discordance was not associated with BMA discharge (crude rate, 2.0% vs 2.3%; aOR 1.02; 95%CI 0.96-1.08; p = 0.557). Compared to patient-physician sex discordance, younger patient age, prior substance use, and prior BMA discharge all had stronger associations with BMA discharge. Patient-physician sex discordance was associated with a small increase in BMA discharge among male patients. This finding may reflect communication gaps, differences in the care provided by male and female physicians, discriminatory attitudes among male patients, or residual confounding. Improved communication and better treatment of pain and opioid withdrawal may reduce BMA discharge.

Health economic evaluation of rehabilitative short-term care : Analysis of the costs of geriatric patients with and without rehabilitative short-term care after inpatient hospital stay

Geriatric patients requiring rehabilitation and admitted to short-term care after an acute inpatient hospital stay seldom receive rehabilitative services later. Rehabilitative short-term care (REKUP) supplements short-term care with rehabilitative measures, aiming to prevent functional restrictions and long-term care. To conduct acost and cost-effectiveness analyses of REKUP and provide data for anationwide rollout. Anon-randomized controlled prospective study was carried out. The intervention group (IG) was paired 1:2 with acontrol group (KG), resulting in the formation of three collectives with follow-up periods of either 30,90 or 180 days (each with IG and KG). Using administrative claims data from the AOK Baden-Württemberg, the mean total costs from the perspective of the health insurance were calculated. Apotential impact of the intervention on costs was analyzed using the difference in differences approach. The analysis comprised 129 patients (IG43; KG86). During the follow-up periods, the IG presented higher rates of rehabilitation and lower rates of long-term care and mortality. Regarding costs, no statistically significant differences were found between the IG and KG in any of the three collectives. For nursing care and medication costs, costs were significantly higher in the follow-up period for the KG, whereas costs for rehabilitation were significantly higher for the IG (p < 0.001). Patients receiving REKUP utilize rehabilitation services more often and have alower likelihood of requiring nursing care or dying with no statistically significant differences in costs. There are potential advantages of REKUP in the target population, which warrant further investigation due to methodological limitations.

Sudden Cardiac Death or Ventricular Arrythmia in Patients Taking Levetiracetam or Oxcarbazepine.

Levetiracetam is a widely used antiseizure medication. Recent concerns have been raised regarding the potential prolongation of the QT interval by levetiracetam and increased risk of sudden cardiac death. This could have profound implications for patient safety and for prescribing practice. This study assessed the potential association of levetiracetam with cardiac outcomes related to QT interval prolongation. We compared outcomes of patients taking levetiracetam with those taking oxcarbazepine as a comparator medication that has not been associated with prolongation of the QT interval. The sample included patients who were newly prescribed levetiracetam or oxcarbazepine from January 31, 2010, to December 31, 2019, using administrative claims data from the OptumLabs Data Warehouse (OLDW). The analysis focused on a combined endpoint of sudden cardiac death or ventricular arrythmia, which are both linked to QT interval prolongation. We used a new user design and selected oxcarbazepine as an active comparator with levetiracetam to minimize bias. We used propensity score weighting to balance the levetiracetam and oxcarbazepine cohorts and then performed weighted Cox regressions to evaluate the association of levetiracetam with the combined endpoint. We identified 104,655 enrollees taking levetiracetam and 39,596 enrollees taking oxcarbazepine. At baseline, enrollees taking levetiracetam were older, more likely to have diagnosed epilepsy, and more likely to have diagnosed comorbidities including hypertension, cerebrovascular disease, and coronary artery disease. In the main analysis, we found no significant difference between levetiracetam and oxcarbazepine in the rate of the combined endpoint for the Cox proportional hazards model (hazard ratio [HR] 0.79, 95% CI 0.42-1.47) or Cox regression with time-varying characteristics (HR 0.78, 95% CI 0.41-1.50). When compared with oxcarbazepine, levetiracetam does not correlate with increased risk of ventricular arrythmia and sudden cardiac death. Our finding does not support the concern for cardiac risk to indicate restriction of levetiracetam use nor the requirement of cardiac monitoring when using it. This study provides Class II evidence that sudden cardiac death and ventricular arrythmia are not more frequent in patients older than 17 years newly prescribed levetiracetam, compared with those prescribed oxcarbazepine.

Women in innovative start-ups and regional inclusiveness: ‘green’ and socially-responsible companies

ABSTRACT We explore the link between women in business decision-making roles, regional gender equality in the workforce and start-up sustainability. We employ unique administrative data on innovative start-ups in Italy and implement a novel crawler-based semantic search of start-ups’ websites to identify ‘green’ firms. Our findings on high-tech innovative start-ups confirm established evidence in the female entrepreneurship literature: greater shares of women in decision-making roles increase the probability of a social focus but decrease the likelihood of science, technology, engineering and mathematics (STEM)-intensive green energy objectives. Instead, the board gender composition does not correlate with non-energy ‘green’ objectives. Non-energy ‘green’ firms emerge more in gender-equal regions, while social start-ups develop where women have fewer employment opportunities.

Characterizing medical patients with delirium: A cohort study comparing ICD-10 codes and a validated chart review method.

Delirium is a major cause of preventable mortality and morbidity in hospitalized adults, but accurately determining rates of delirium remains a challenge. To characterize and compare medical inpatients identified as having delirium using two common methods, administrative data and retrospective chart review. We conducted a retrospective study of 3881 randomly selected internal medicine hospital admissions from six acute care hospitals in Toronto and Mississauga, Ontario, Canada. Delirium status was determined using ICD-10-CA codes from hospital administrative data and through a previously validated chart review method. Baseline sociodemographic and clinical characteristics, processes of care and outcomes were compared across those without delirium in hospital and those with delirium as determined by administrative data and chart review. Delirium was identified in 6.3% of admissions by ICD-10-CA codes compared to 25.7% by chart review. Using chart review as the reference standard, ICD-10-CA codes for delirium had sensitivity 24.1% (95%CI: 21.5-26.8%), specificity 99.8% (95%CI: 99.5-99.9%), positive predictive value 97.6% (95%CI: 94.6-98.9%), and negative predictive value 79.2% (95%CI: 78.6-79.7%). Age over 80, male gender, and Charlson comorbidity index greater than 2 were associated with misclassification of delirium. Inpatient mortality and median costs of care were greater in patients determined to have delirium by ICD-10-CA codes (5.8% greater mortality, 95% CI: 2.0-9.5 and $6824 greater cost, 95%CI: 4713-9264) and by chart review (11.9% greater mortality, 95%CI: 9.5-14.2% and $4967 greater cost, 95%CI: 4415-5701), compared to patients without delirium. Administrative data are specific but highly insensitive, missing most cases of delirium in hospital. Mortality and costs of care were greater for both the delirium cases that were detected and missed by administrative data. Better methods of routinely measuring delirium in hospital are needed.

Local area variation in health visiting contacts across England for children under age 5: a cross-sectional analysis of administrative data in England 2018-2020

BackgroundThe health visiting service in England leads the government's Healthy Child Programme (HCP) for children under five years. Local authorities and their provider partners deliver this service differently across England. ObjectiveTo describe local authority variation in the delivery of health visiting to children under five years in England (2018-2020). MethodsWe used publicly available statistics on mandated health visiting contacts, and administrative data from the Community Services Dataset (CSDS) on duration, location, and medium of contacts. We mapped population coverage of mandated contacts (new birth visit, 6-8-week review, one-year review, and 2-2½ -year review) and described the frequency and characteristics of mandated and additional contacts across local authorities. ResultsBased on publicly available data, almost all eligible children received their new birth visit, 6-8-week review and one-year review (89%-99%), with substantial variation across local authorities in children receiving he 2-2½ -year review: median 81%, range 33%-98%. Based on CSDS, 80% of local authorities (n=46/57) delivered more additional than mandated contacts: a median of 1.6 additional contacts (range: 0.1-8.5) were delivered for each mandated contact. There was also significant variation in the duration of contacts and the percentage of contacts delivered face-to-face and at home. ConclusionsDespite decreases in funding and workforce since 2015, in 2018--2020, health visiting teams reached nearly all babies and most children face-to-face via mandated contacts, and conducted over one and a half times the number of additional contacts relative to mandated contacts, with variation between local areas. This represents a significant public health infrastructure to support the health and development of babies and children and the wellbeing of their families in the critical period before school. Our study highlights the importance of taking into account additional contacts. Further work is needed to understand variation, including in the way additional contacts are used.

Predictors of Adverse Drug Reaction-Related Hospitalisations Among People with Dementia: A Retrospective Case-Control Study.

Adverse drug reactions (ADRs) are common among people with dementia; however, little is known about the magnitude and predictors associated with ADR-related hospitalisation among these individuals. This study aimed to determine the magnitude, types, drugs implicated and predictors of ADRs associated with hospitalisation among people with dementia. This retrospective case-control study analysed medical records of individuals aged ≥ 65 years with dementia admitted to major public hospitals in Tasmania, Australia, from July 2010 to July 2021. Adverse drug reactions and implicated drugs were identified using administrative data and cross-checked with hospital medical records, with consensus reached among the research team. Of the 7928 people admitted to hospital at least once within the study period, 1876 (23.7%) experienced at least one ADR-related hospitalisation. Of these, 300 case patients with 311 ADRs and 300 control patients were randomly selected. The most common types of ADRs were renal (acute kidney injury; AKI) (36.0%), followed by neuropsychiatric (17.6%), cardiovascular (16.0%) and haematological (13.1%). Diuretics, renin-angiotensin system (RAS) inhibitors and anti-thrombotics constituted the main implicated drug classes. The ADR-related hospitalisation was associated with: chronic kidney disease (CKD) (OR 8.00, 95% CI 2.63-24.28, p < 0.001), Australian-born (OR 1.62, 95% CI 1.08-2.43, p = 0.019), hypertension (OR 1.48, 95% CI 1.01-2.17, p = 0.044) and the number of medicines (OR 1.06, 95% CI 1.00-1.12, p = 0.022). Potentially inappropriate medication use and anticholinergic burden did not predict ADR-related hospitalisation. These predictors could help identify the individuals at the highest risk and enable targeted interventions to be designed.

The role of bipolar disorder and family wealth in choosing creative occupations

Research in psychology and medicine has linked mental health disorders, and particularly bipolar disorder (BD), to employment in creative professions. Little is known, however, about the mechanisms for this link, which could be due to biology (primarily through a person’s genes) or environmental (through socioeconomic status). Using administrative data on mental health diagnoses and occupations for the population of Denmark, we find that people with BD are more likely to be musicians than the population, but less likely to hold other creative jobs. Yet, we also show that healthy siblings of people with BD are significantly more likely to work in creative professions. Notably, people from wealthy families are consistently more likely to work in creative professions, and access to family wealth amplifies the likelihood that siblings of people with BD pursue creative occupations. Nevertheless, family wealth explains only a small share of the correlation between BD and creative employment.

United States Veterans with comorbid chronic fatigue syndrome and chronic pain: do women differ from men?

ABSTRACT Objective: Examine group differences among women and men United States Veterans with chronic fatigue syndrome (CFS) and chronic pain (CP). Method: Administrative data from the VA Corporate Data Warehouse were derived from a cohort that met criteria for CP in 2018. Differences across sociodemographic characteristics, psychiatric comorbidities, and pain-related healthcare utilization were examined. Analyses included group comparisons (chi-square tests for independence or independent samples t-tests) and effect size differences (Cohen’s d or Cramer’s V). Results: The period prevalence for comorbid CFS and CP was greater among women Veterans (1.20% vs. 0.61%). Women Veterans with CFS and CP were younger (d = 0.80), more likely to be Black, and less likely to be White (V = 0.13). Women Veterans with CFS and CP were more likely to have fibromyalgia (V = 0.23), headache (V = 0.24), or urogenital, pelvic, and menstrual pain (V = 0.14), and less likely to have neuropathy relative to men; they were also more likely to have depression (V = 0.15) or anxiety (V = 0.13). Conclusions: Comorbid CFS and CP are more prevalent among women Veterans relative to men. Group differences in pain clusters and psychiatric comorbidities may have implications for clinical offerings to women Veterans with these conditions.

MRI use leading up to total knee arthroplasty: a retrospective cohort study.

Demand for total knee arthroplasty (TKA) is increasing as it remains the gold-standard treatment for end-stage osteoarthritis (OA) of the knee. Magnetic resonance imaging (MRI) scans of the knee are not indicated for diagnosing knee OA and represent a possible delay to orthopaedic surgeon referral and unnecessary expenditure. The purpose of this study was to determine the proportion of patients who underwent an MRI in the two years prior to their primary TKA for OA and determine patient and physician associations with increased MRI usage. This is a population-based cohort study using administrative data from Ontario, Canada. All patients over 40years old who underwent their first primary TKA between April 1, 2008, and March 31, 2019, were included. Statistical analyses were performed using SAS and included the Cochran-Armitage test for trend of MRI prior to surgery. A predictive multivariable regression model was used to determine features correlated to receiving an MRI. There were 194,989 eligible first-time TKA recipients, of which 38,244 (19.6%) received an MRI in the two years prior to their surgery. The majority of these (69.6%) were ordered by primary care physicians. Patients who received an MRI were younger, had fewer comorbidities and were more affluent than patients who did not (p < 0.001). MRI use prior to TKA increased from 2008 to 2018 (p < 0.001). Despite MRIs rarely being indicated for the work-up of end-stage OA, nearly one in five patients have an MRI in the two years prior to their TKA. This may be increasing healthcare expenditure and surgical wait-times.

How does digital technology administrative penalty affect big data technology innovation: evidence from China

Administrative regulation is an essential institutional arrangement for governing innovation in big data technology. Administrative penalties are one of its main methods. It is worth investigating whether administrative penalties can serve as an institutional safeguard to promote technological innovation. This study collects data on digital technology administrative penalty (DTAP) and big data technology innovation (BDTI) from 2008 to 2020 in 281 cities in China and empirically evaluates the impact and mechanism of DTAP on BDTI. The findings suggest that the normative impact of the DTAP system can foster a conducive business ecosystem for big data innovation. The incentive mechanism motivates firms to increase their long-term investment in technological innovation, while the deterrence mechanism ensures the existence of a regulated competitive market. These mechanisms play a crucial role in facilitating BDTI. Mechanism tests show that the DTAP has the potential to promote novel business models in the digital economy and accelerate progress in industrial digitalisation, which in turn promotes innovation in big data technologies. The impact of the DTAP on promoting BDTI is diverse, with a significantly greater impact in the first-tier cities. DTAP plays a more important role in fostering BDTI in places where the digital factor-driven industry is still in its early stages of development.

Mode of delivery and birth outcomes before and during COVID-19 -A population-based study in Ontario, Canada.

There is lack of clarity on whether pregnancies during COVID-19 resulted in poorer mode of delivery and birth outcomes in Ontario, Canada. We aimed to compare mode of delivery (C-section), birth (low birthweight, preterm birth, NICU admission), and health services use (HSU, hospitalizations, ED visits, physician visits) outcomes in pregnant Ontario women before and during COVID-19 (pandemic periods). We further stratified for pre-existing chronic diseases (asthma, eczema, allergic rhinitis, diabetes, hypertension). Deliveries before (Jun 2018-Feb 2020) and during (Jul 2020-Mar 2022) pandemic were from health administrative data. We used multivariable logistic regression analyses to estimate adjusted odds ratios (aOR) of delivery and birth outcomes, and negative binomial regression for adjusted rate ratios (aRR) of HSU. We compared outcomes between pre-pandemic and pandemic periods. Possible interactions between study periods and covariates were also examined. 323,359 deliveries were included (50% during pandemic). One in 5 (18.3%) women who delivered during the pandemic had not received any COVID-19 vaccine, while one in 20 women (5.2%) lab-tested positive for COVID-19. The odds of C-section delivery during the pandemic was 9% higher (aOR = 1.09, 95% CI: 1.08-1.11) than pre-pandemic. The odds of preterm birth and NICU admission were 15% (aOR = 0.85, 95% CI: 0.82-0.87) and 10% lower (aOR = 0.90, 95% CI: 0.88-0.92), respectively, during COVID-19. There was a 17% reduction in ED visits but a 16% increase in physician visits during the pandemic (aRR = 0.83, 95% CI: 0.81-0.84 and aRR = 1.16, 95% CI: 1.16-1.17, respectively). These aORs and aRRs were significantly higher in women with pre-existing chronic conditions. During the pandemic, healthcare utilization, especially ED visits (aRR = 0.83), in pregnant women was lower compared to before. Ensuring ongoing prenatal care during the pandemic may reduce risks of adverse mode of delivery and the need for acute care during pregnancy.

Prenatal opioid exposure by likelihood of exposure and risk to prenatal development: Medicaid-covered births in Wisconsin, 2010-2019.

Prenatal opioid exposure is an established public health problem, in particular among Medicaid-covered births. Yet, existing prevalence rates are plausibly underestimated. We leverage extensive linked longitudinal administrative data for all Medicaid-covered live births in Wisconsin from 2010 to 2019 to estimate a range of prevalence rates using an innovative strategy that jointly accounts for both likelihood of exposure and potential risk to prenatal development. We find that 20.8% of infants may have been prenatally exposed to opioids, with 1.7% diagnosed with neonatal abstinence syndrome and an additional 1.2% having a high combined likelihood of exposure and potential risk to prenatal development, 2.6% a moderate combined likelihood and risk, and 15.3% a low or uncertain combined likelihood and risk. We assess improvements in prevalence estimates based on our nuanced classification relative to those of prior studies. Our strategy could be broadly used to quantify the scope of the opioid crisis for pregnant populations, target interventions, and promote child health and development.

What Explains Inequalities in Telehealth Utilization Among North Carolina Medicaid Beneficiaries?

Background: Increased availability of telehealth can improve access to health care. However, there is evidence of persistent disparities in telehealth usage, as well as among people from minoritized racial and ethnic groups and rural residents. The objective of our work was to explore the degree to which disparities in telehealth use for behavioral health (BH) and musculoskeletal (MSK) related services during the COVID-19 pandemic are explained by observed beneficiary- and area-level characteristics. Methods: Using North Carolina Medicaid claims data of Medicaid beneficiaries with BH or MSK conditions, we apply nonlinear regression-based decomposition analysis-based models developed by Kitagawa, Oaxaca, and Blinder to determine which observed variables are associated with racial, ethnic, and rural inequalities in telehealth usage. Results: In the BH cohort, we found statistically significant differences in telehealth usage by race in the adult population, and by race, Hispanic ethnicity, and rurality in the pediatric population. In the MSK cohort, we found significant inequities by Hispanic ethnicity and rurality among adults, and by race and rurality among children. Inequalities in telehealth use between groups were small, ranging from 0.7 percentage points between urban and rural adults with MSK conditions to 3.8 percentage points between white adults and people of color among those with BH conditions. Overall, we found that racial and ethnic inequalities in telehealth use are not well explained by the observed variables in our data. Rural disparities in telehealth use are better explained by observed variables, particularly area-level broadband internet use. Conclusions: For inequalities between rural and urban residents, our analysis provides observational evidence that infrastructure such as broadband internet access is an important driver of differences in telehealth use. For racial and ethnic inequalities, the pathways may be more complex and difficult to measure, particularly when relying on administrative data sources in place of more detailed data on individual-level socioeconomic factors.

The impact of terminating cost-sharing reductions payments on health insurance plan choices

The Trump administration terminated cost-sharing reductions (CSRs) payments to health insurers in 2017, while still required insurers to provide CSRs to eligible enrollees in the Marketplace. Marketplace administration data reveals that, in response to this termination, insurers raised premiums to compensate for their loss. Consequently, premium increases led to more advanced premium tax credits for enrollees in the Marketplace. To investigate the impact of CSRs payment termination on low-income consumer insurance plan choices, I leverage variations in state price regulations and employed a difference-in-differences design. In a robustness analysis, I utilized differences in county income distributions to examine the effects of the termination on insurance choices. The results indicate that after the termination, more low-income enrollees opted for cheaper bronze plans, and fewer chose silver plans. These results suggest that alterations in subsidy channels may inadvertently encourage low-income individuals to purchase less expensive health insurance plans, highlighting an unintended consequence of the termination of cost-sharing subsidies.

The impact of health impairments on employment entry and the quality of employment among basic income support recipients in Germany

Abstract Policies for recipients of basic income support for jobseekers in Germany focus on activation and quick labour market integration. Yet, the majority of benefit recipients report severe health impairments. Against this background, the article investigates implications of health impairments for benefit recipients’ jobcentre relationship and employment opportunities. The analyses show that 63 per cent of non-employed benefit recipients report health restrictions on their employment capabilities, 51 per cent report severe health impairments, and 25 per cent that they cannot work at all. The most frequent types of health impairments are musculoskeletal and mental health impairments. Health impairments significantly reduce entry rates into socially insured employment, but do not seem to inhibit taking up uninsured minijobs. Counselling frequency increases job entry rates for benefit recipients without health impairments in the short-term. For those with health impairments, no short-term effects are found over a one-year follow-up period. Policy responses could include a more explicit acknowledgement of health impairments as a central issue for benefit recipients. Greater investments in rehabilitation and subsidised employment could be part of a strategy to improve opportunities for benefit recipients with health impairments to find better-quality (part-time) employment instead of uninsured minijobs. The analyses are based on linked longitudinal PASS survey and administrative data.

Health Care Utilization During the First 5½ Years of Authorized Service by Transgender U.S. Service Members (2016-2021).

Purpose: This study examined the utilization of gender-affirming health care by active-duty service members during the initial 5½ years that transgender and gender-diverse (TGD) individuals were authorized to serve in the U.S. military. The aim of this study was to inform policy discussions regarding inclusion of TGD individuals in the military. Methods: We conducted a retrospective cohort study using administrative health care data from the Military Data Repository (MDR). We assessed the association of demographic factors with the utilization of gender-affirming medical and surgical care. We calculated the incidence rate of initial TGD-related encounters and new prescriptions for gender-affirming hormones among Defense Department Service members between July 2016 and December 2021. Results: We identified 2481 service members with an initial health care encounter with an associated TGD-related diagnosis. More than half (53%) of these service members started gender-affirming hormones, and 14% underwent gender-affirming surgery. Mastectomies and hysterectomies accounted for more than 70% of surgeries. Service members made 3.22 initial encounters with a TGD-related diagnosis per 10,000 service members per year, with an overrepresentation of service members who were designated female in the MDR (8.62), junior enlisted (4.98), and young (4.64). Individuals designated female in the MRD comprised 17% of all service members but accounted for 46% of initial encounters, 51% of new prescriptions, and 73% of surgeries. Conclusion: The study revealed a higher-than-expected number of service members seeking gender-affirming care, particularly among service members designated female in the MDR. Military Health System clinicians provided most of this care, which may mitigate the cost of delivering this essential medical care.

Administrative Management Data Anomaly Access Detection Method, Based on 6G IoT

Administrative Management Data Anomaly Access Detection Method, Based on 6G IoT

Care cascades following low-value cervical cancer screening in dually enrolled Veterans.

Veterans dually enrolled in the Veterans Health Administration (VA) and Medicare commonly experience downstream services as part of a care cascade after an initial low-value service. Our objective was to characterize the frequency and cost of low-value cervical cancer screening and subsequent care cascades among Veterans dually enrolled in VA and Medicare. This retrospective cohort study used VA and Medicare administrative data from fiscal years 2015 to 2019. The study cohort was comprised of female Veterans aged >65 years and at low risk of cervical cancer who were dually enrolled in VA and Medicare. Within this cohort, we compared differences in the rates and costs of cascade services related to low-value cervical cancer screening for Veterans who received and did not receive screening in FY2018, adjusting for baseline patient- and facility-level covariates using inverse probability of treatment weighting. Among 20,972 cohort-eligible Veterans, 494 (2.4%) underwent low-value cervical cancer screening with 301 (60.9%) initial screens occurring in VA and 193 (39%) occurring in Medicare. Veterans who were screened experienced an additional 26.7 (95% CI, 16.4-37.0) cascade services per 100 Veterans compared to those who were not screened, contributing to $2919.4 (95% CI, -265 to 6104.7) per 100 Veterans in excess costs. Care cascades consisted predominantly of subsequent cervical cancer screening procedures and related outpatient visits with low rates of invasive procedures and occurred in both VA and Medicare. Veterans dually enrolled in VA and Medicare commonly receive related downstream tests and visits as part of care cascades following low-value cervical cancer screening. Our findings demonstrate that to fully capture the extent to which individuals are subject to low-value care, it is important to examine downstream care stemming from initial low-value services across all systems from which individuals receive care.

Predicting unplanned hospitalisations in older adults using routinely recorded general practice data.

Unplanned hospitalisations represent a hazardous event for older persons. Timely identification of high-risk individuals using a prediction tool may facilitate preventive interventions. To develop and validate an easy-to-use prediction model for unplanned hospitalisations in community-dwelling older adults using readily available data to allow rapid bedside assessment by general practitioners. Retrospective study using general practice electronic health records of 243,129 community-dwelling adults aged ≥65 years linked with national administrative data. The dataset was geographically split into a development (58.7%) and validation (41.3%) sample to predict unplanned hospitalisations within 6 months. We evaluated the performance of three different models with increasingly smaller selections of candidate predictors (i.e. optimal, readily-available and easy-to-use model, respectively). We used logistic regression with backward selection for model development. The models were validated internally and externally. We assessed predictive performance by area under the curve (AUC) and calibration plots. In both samples, 7.6% had at least one unplanned hospitalisation within 6 months. The discriminative ability of the three models was comparable and remained stable after geographic validation. The easy-to-use model included age, sex, prior hospitalisations, pulmonary emphysema, heart failure and polypharmacy. Its discriminative ability after validation was AUC 0.72 [95% confidence interval: 0.72-0.71]. Calibration plots showed good calibration. Our models showed satisfactory predictive ability. Reducing the number of predictors and geographic validation did not impact predictive performance, demonstrating the robustness of the model. We developed an easy-to-use tool that may assist general practitioners in decision-making and targeted preventive interventions.