Cancer treatment and screening

Einleitung / Introduction

Treatment of localized prostate cancer is well established and consists primarily of expectant management (active surveillance) and definitive treatment (radical prostatectomy or radiation). Yet, the evidence to date of their comparative effectiveness relies on randomized controlled trials (RCTs). Small sample sizes and self-selection of trial participants limit the generalizability of RCTs to common care settings and the general population. Past trials have enrolled few members of racial-minority and rural populations, even though these groups often are diagnosed with more aggressive or advanced forms of prostate cancer and develop it more frequently.

Methode / Method

We tested whether clinical and claims data from a broadly representative cross-section of the US population of prostate cancer patients could be used to estimate the causal effect of expectant management versus definitive treatment on prostate cancer outcomes by applying a regression-discontinuity (RD) design. Our RD design exploits the fact that prostate cancer risk classification, which is based on several biomarkers such as tumor size and spread, Gleason score and prostate-specific antigen (PSA) value, is discrete. In contrast, prostate cancer progression is arguably a continuous, biological process that does not change discontinuously at risk-classification thresholds. Thus, we tested the hypothesis that patients with practically identical severity of prostate cancer on either side of the risk classification threshold would choose different treatment purely due to their assignments to different risk groups. We used records from the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database of prostate cancer patients aged 65 to 84 years who were diagnosed between 2004 and 2017.

Ergebnisse / Results

In the age group 65 to 74 years, patients with PSA = 20 were 5.3 percentage points (p value = 0.008) less likely to choose expectant management within six months of diagnosis than otherwise identical patients whose PSA value was 19.9. As 21.3% of patients with PSA = 18 or 19 chose expectant management, this drop represents a 24.9% decrease that is attributable to the classification as high risk. In the age group 75 to 84 years, patients with PSA = 20 were 9.4 pp (p value = 0.009, -16.9%) less likely to choose expectant management. There were no statistically significant differences between the two risk groups at the critical threshold in cancer severity, socio-demographic characteristics, or health care use.

Zusammenfassung / Conclusion

Data from a universal public insurance program can be used to uncover the causal effect of treatment of localized prostate cancer on outcomes. We plan to present results for health care quality, use, and cost, treatment complications, disease recurrence, and survival.

Authors

Lorens Helmchen, The George Washington University

Robert Kaestner, The University of Chicago

Einleitung / Introduction

Introduction The burden of disease in cancer is increasing despite significant advancements in medical technology. Germany introduced accreditation of local provider networks as organ cancer centers to assure quality and to develop oncological care structures. Accreditation, which occurs earlier in districts with higher innovation activity to develop new cancer treatments, is associated with improved treatment outcomes. Uneven accreditation uptake across regions may limit access to high quality care for certain patient populations. In this study, we evaluate the effect of the accreditation of specialized cancer care on patient survival.

Methode / Method

Method We use cancer registry data from the years 2000 until 2018 on the patient level to predict survival time and probabilities for different patient populations. We aggregate the data on the regional level for 401 districts in Germany and combine it with accreditation, socioeconomic and innovation activity data. In a difference-in-difference approach, we estimate the causal effect of accreditation on cancer survival after diagnosis for seven cancer sites.

Ergebnisse / Results

Results The patient data contains 5.19 M patients of which 24% had access to at least one accredited center related to their cancer type in their district at time of diagnosis. Average probability to survive one and five years after diagnosis was 0.71 and 0.45, respectively. After adjusting for patient characteristics and regional innovation activity, linear regression with cancer type fixed effects suggests that patients who lived in districts with at least one accredited center related to their cancer type had a 11.8 (p<0.001) percentage points higher probability to survive at least five years after diagnosis. The probability increased by 13.7 (p<0.001) and 15.5 (p<0.001) percentage points in districts in which at least one third and two thirds of the hospitals respectively were accredited. Patients in districts with any publication activity in the diagnosis year had a 11.7 (p<0.001) percentage points higher probability to survive five years.

Zusammenfassung / Conclusion

Contribution We contribute by identifying a causal effect of accreditation of specialized cancer care on survival of cancer patients and accounting for innovation activities as a major driver of improved survival rates of cancer patients. The results inform and develop the strategies of provider organizations and health policy.

Authors

Tim Brand, Universität Duisburg-Essen

Katharina Blankart, Universität Duisburg-Essen

Einleitung / Introduction

Forgone Care, i.e. not seeking care when needed, is a long-standing problem with multiple causes. Forgoing preventive screenings, such as mammography, not only carry the risk of delayed diagnosis of serious disease, but can also increase overall morbidity and all-cause mortality, long-term complications, and healthcare costs. It is thus the aim of our study to provide a claims data analysis of the impact of forgone care on health outcomes during the Covid-19 pandemic in Germany.

Methode / Method

The analysis is based on anonymized data from multiple German sickness funds accessed through GWQ ServicePlus AG. First, we assessed the existence of forgone care by comparing mammography screenings before (2014-2019) and during the pandemic (2020-2022). We differentiate between screenings that are invitation-based (IBS; reimbursement code 01750) and referral-based screenings (RBS; reimbursement code 34270). Second, we investigate the link between the reduction in mammography screening frequencies/rates and health outcomes and costs. To do so, we compare women who are eligible for screening before (2017) and during (2020) the pandemic and have no prior history of cancer within 3 years. We measure all-cause mortality rates, the period between the first screening and the start of treatment, changes in the distribution of cancer treatment, and costs. We use entropy balancing to adjust for socio-demographic characteristics, such as age and sex, and morbidity, as measured by the Elixhauser index. Thereafter we apply weighted regression models for health outcomes and DiD regression for costs. Subgroup analysis of those who actually participate in the screenings will be conducted.

Ergebnisse / Results

Preliminary results indicate a decrease in mammography screening of up to 44% during the pandemic. From 2014 to 2019, an average of 6990 IBS and 6874 RBS were performed annually. Overall, the pre-pandemic screening trend remains stable, although seasonal fluctuations, especially towards the end of year, occur. However, in 2020, only 2454 IBS were performed annually, while RBS only decreased slightly to 5999. This trend continues in 2021, particularly for IBS (2586), while RBS is approaching the pre-pandemic figures (6039), but is still below the average (2014-2019). The analyses of how forgone screenings impact health outcomes and costs has yet to be conducted.

Zusammenfassung / Conclusion

As breast cancer is one of the most common cancers in women, we expect that missed mammograms will be of great importance for health outcomes. As no catch-up effect has been observed so far, it is important to understand the consequences of the (continuous) reduction in prevention activities on health outcomes and costs. Forgone care remains a pressing issue, as people forgo necessary medical treatment for various reasons.

Authors

Annabelle Röpcke, Hamburg Center for Health Economics, Universität Hamburg

Tom Stargardt, Hamburg Center for Health Economics, Universität Hamburg

André Hajek, Universitätsklinikum Hamburg

Lea Grotenrath, GWQ-ServicePlus AG