The EuHEA Seminar Series aims to foster exchange between health economists across different countries and institutions and present cutting-edge research in all areas of health economics. A Scientific Committee chaired by Geir Godager (University of Oslo) and Pedro Pita Barros (Universidade Nova de Lisboa) coordinated the series in spring 2021.
Please note that the seminars were not recorded.
11 February 2021, 14:00-15:00 (CET)
Objectives: The study aims at analysing possible factors affecting women mental wellbeing in the aftermath of the COVID19 first wave. Comparing alternative measures of mental health, we focus on two possible categories of drivers: present concerns and future expectations. The former are defined as factors already present in the experience of the respondent (intensity of the outbreak, changes in unemployment status, stay-at-home restrictions), the latter as speculations on future events (future changes in access to care or employment conditions).
Methods: We run a cross-sectional survey in July 2020 on more than 4,000 women aged 20-65 resident in Italy. The selected group is representative of the real population considered on a geographic and demographic basis. Main outcomes are the level of self-assessed mental distress and concerns, while we alternatively study the use of medications and the self-assessed health. The analysis is performed by means of an OLS estimation, where sets of individual covariates are included (socio-demographics, household characteristics, personal and partner’s employment, present concerns, expectations). We provide further insights on the role of expectations performing a heterogeneity analysis over alternative indexes of gender stereotypes to test for the role of a potential mismatch between women’s aspirations and societal expectations.
Results: Present concerns play a minor role compared to expectations on the future. Beside a strong gradient by age group, with younger women being the most affected, main explanatory factors are the fear to lose the job (own/partner) and negative expectations about labor market and access to care. Interestingly, concerns about own employment status in the future prevail over similar concerns for the partner’s condition, meaning that the worst effects on mental health are related to the fear to lose the social status linked to being employed, rather than the mere expected reduction in household income. Other relevant but minor factors are educational attainments, having (underage) children, practicing remote work. More conservative gender norms increase the probability to report a poor mental health status but negatively interact with expectations, with women living in context with stronger stereotypes partially compensating the effect of negative expectations on their mental wellbeing.
Discussion: The unexpected pandemic situation has large effects on the mental wellbeing of individuals, at least for the specific population group of women who are more exposed to adverse consequences in the labor market and generally more affected by mental health distress. Policy makers should consider that expectations seem to be the main responsible for adverse mental health outcomes, defining adequate interventions that go beyond financial supports or the removal of COVID19 restrictions. Interestingly, gender stereotypes end up acting as a coping mechanism to deal with stress about future working life.
18 February 2021, 14:00-15:00 (CET)
Objectives: This study aims to identify the factors associated with 1) lockdown support during the 2nd lockdown in France and plans for end-of-year holidays’ celebrations and 2) the acceptability and the use of the contact-tracing application ‘TousAntiCovid’.
Methods: Between November 20th and 23rd 2020, a cross-sectional study was conducted among a representative sample of the French population. Factors associated with lockdown support are estimated using an ordered logistic regression. Factors associated with preferences regarding the level of restrictions during the end-of year holidays and with plans for holidays’ celebrations are estimated using multinomial logistic models. Factors associated with the current use and the intention to use the ‘TousAntiCovid’ application are estimated using a logistic model and ordered logistic model respectively. Independent variables of interest include COVID-19 perceived threat, perceived benefits/costs of lockdown and of digital contact-tracing, trust in the government, health literacy, time and risk preferences and endorsement of COVID-19 conspiracy beliefs.
Results: Lockdown support is positively correlated with trust in the government and perceived lockdown efficacy while it is negatively correlated with the endorsement of COVID-19 conspiracy beliefs. Respondents perceiving the health impacts of COVID-19 as very serious and the lockdown as effective are less likely to reject the implementation of a lockdown during the end-of-year holidays. Less support for a strict lockdown during the holidays is found among respondents judging the lockdown as too costly. Respondents perceiving the health impacts of COVID-19 as serious, those trusting the government and those with a higher health information appraisal score are less likely to celebrate holidays as usual. Future-oriented respondents and those with a lower willingness to take health risks are also more likely to change their celebrations’ plans compared to previous years. The use and intention to use the ‘TousAntiCovid’ application are positively correlated with the perceived efficacy and the perceived data privacy of the application while they are negatively correlated with the COVID-19 conspiracy score.
Discussion: Public communication should fight misinformation on COVID-19 that reduce the support to containment measures and decrease the likelihood to use the ‘TousAntiCovid’ application. As lockdown measures might be reinstated in France, public authorities should also communicate around the effectiveness of such measures and reinforce the associated financial aid programs in order to enhance lockdown acceptance among the population. As ‘TousAntiCovid’ is a key part in the French ‘test, trace, isolate’ strategy, public authorities should further communicate on the utility of digital contact-tracing and reinsure the public on the security of data collected in order to increase the adoption of the application.
25 February 2021, 14:00-15:00 (CET)
Objectives: The opportunity to consult a doctor online via video calls or chats is a recent phenomenon in health care. Knowledge is scarce regarding how the availability of online consultations affect individuals’ consumption of regular health care services. In this paper, we causally examine to which extent online consultations replace in-person doctor consultations in Swedish primary care. In Sweden, online consultations was a marginal phenomenon before 2016, when a few private companies realized that they could exploit an institutional feature granting them public funding on a fee-for-service basis. In 2018, these companies supplied 5% of all doctor consultations in primary care. If online consultations fully substitute for in-person visits, the effect on health care spending is only a matter of comparing unit costs. In practice, the degree of substitution is limited by several factors. First, some cases require a physical examination, and thus one additional visit. Second, the greater convenience of online contacts might increase the total demand. Third, the online doctors might specialise on patients with mild and transitory conditions, who might not have been treated at all by regular primary care.
Methods: We use administrative data for 2013-2018 covering in-person and online consultations by 19-20 year olds in the two most populous Swedish regions. The major challenge to identification is that unobserved transitory health problems might simultaneously affect the individual’s consumption of online and in-person care. Of the few existing studies on the subject, ours is the first to account for such heterogeneity. We do so by exploiting exogenous variation in patient fees for online visits facing patients at their 20th birthday in a regression discountinuity design. To limit concerns for other confounding factors at the 20th birthday, we compare the jumps in in-person visits around the 20th birthday of birth cohorts that reached the age limit before and after the online market emerged. We use the differential discontinuity as an instrumental variable for the number of online consultations in a fuzzy RD.
Results [preliminary]: Our main estimate suggest that roughly every other online visit replaces an in-person visit. Heterogeneity analyses suggest that the degree of substitution is larger for men than for women.
Discussion: The net effect in these age groups and in these regions is to increase the number of doctor consultations in primary care. This is in line with previous evidence from two US studies and a study from another Swedish region. Online consultations decrease patients’ private costs for seeking care, as no time is spent travelling to the practice or in the waiting room. This greater convenience appears to come at the cost of increasing the moral hazard problem of health insurance.
4 March 2021, 14:00-15:00 (CET)
Objectives: Decision makers frequently use cost-sharing to alleviate pressure on public healthcare budgets. Apart from generating revenue directly, cost-sharing is a means to influence and steer the behaviour of patients to control demand for healthcare services and thereby address moral hazard. The effect of cost-sharing on demand for healthcare services has been heavily studied in the literature, but researchers often apply a macro-perspective on these issues, opening the door for the fallacy of assuming uniform demand reactions across a spectrum of different healthcare services. The aim of this article is to estimate price elasticities of a variety of healthcare services to highlight how they depend on urgency and price.
Methods: We utilise a dataset of pseudonymised longitudinal patient-level data on healthcare service consumption between Q2-2015 and Q2-2017 of three different sickness funds in Austria covering 1,035,177 patients. We estimate the price elasticity of a set of 11 healthcare services differing in urgency and price. We combine matching via entropy balancing and difference-in-differences estimation in a two-stage study design following a reduction in the co-insurance rate by one of the sickness funds from 20% to 10% in Q2-2016. We further test the robustness of our result using different frequencies on the dependent variable and placebo regression.
Results: The reduction of the co-insurance rate led to a small increase in demand for routine ECGs (+1.5%) and a negligible increase for electromyography (+0.1%) over the whole post-treatment period. Only the effect for routine ECG is statistically significant and robust to our sensitivity analyses. For the nine other healthcare services, pre-trends fail the necessary conditions for a difference-in-differences framework.
Discussion: Our results show that price elasticities of different healthcare services depend on their urgency and costs and cast a new light on previous empirical evidence on price elasticity of healthcare services derived without differentiation between services. Routine ECGs and electromyography are two comparatively expensive healthcare services in the outpatient sector. But whereas routine ECGs are often performed during a health check-up and can easily be postponed by patients, electromyography is more urgent and patients do not have discretion over the timing of the healthcare service consumption. For healthcare services that are urgent, low cost or both, we do not find evidence that a change in co-insurance rate affects demand. A limitation to our study is that some of the healthcare services are not frequently consumed and may be prone to distortions by regional or seasonal fluctuations which may cause deviations in pre-trends. In combination with a small effect size, this likely contributes to the comparatively low statistical significance of the findings.
11 March 2021, 14:00-15:00 (CET)
Objectives: Switching costs and persistent patient preferences generate demand inertia and link current and future choices of hospital. If the choices patients make are intertemporally linked, these choices will be affected by whether patients anticipate the future, as well as the degree of sophistication of their foresight. Motivated by this observation, we investigate the effect of patient expectations (whether and how patients anticipate the future) on quality provision.
Methods: We develop a two-period spatial model of hospital competition, where demand inertia results from both persistent patient preferences and switching costs. Hospitals are assumed to be "motivated", and we allow for either cost substitutability or complementarity between quality and output. We consider three types of expectations. Myopic patients choose a hospital based on current variables alone; forward-looking but naive patients take the future into account, but assume that quality remains constant; and forward-looking and rational patients correctly foresee the evolution of quality.
Results: We first show that patient expectations affect quality provision through the responsiveness of demand to quality, with higher responsiveness leading to higher provision. Then, we rank quality provision and show that it is higher under naive than myopic expectations, while quality under rational expectations may be highest, lowest, or lie in between. We also show that only under rational expectations may quality be lower than in the counterfactual case of a market without inertia and policies aimed at reducing switching costs beneficial.
Discussion: Discussions of the role of rationality commonly focus on the idea that deviations from fully rational behaviour make consumers act not in their best interest and that firms may find it beneficial to exploit those deviations. Our results indicate that the reverse might as well hold in hospital markets, as our quality ranking also holds for patients’ health gains. When a unilateral increase in current quality yields a reduction in the future quality difference, rational patients foresee its effect on their total expected utility. They are, thus, less sensitive to quality than they would be if they ignored the future. Myopic and naive patients, differently, are oblivious to the future quality reduction and hence overestimate the impact of the current quality increase on their total utility, which leads to higher demand responsiveness and induces hospitals to invest in quality. In this case, therefore, the departures from rationality insulate patients from inferior quality provision by hindering the hospitals’ ability to exploit the otherwise lower demand responsiveness
18 March 2021, 14:00-15:00 (CET)
Objectives: Patients who are themselves experts have been found to receive care that is systematically different from care provided to non-expert patients. However, the current literature has been unable to ascertain whether the differences are due to expert patients sending less noisy signals about their preferences or health state than non-experts (statistical discrimination theory) or whether experts use their informational advantage to demand better care than non-expert patients (agency discrimination theory). We investigate the extent to which care provided to medically trained mothers is more likely to bypass clinical guidelines intended to ration access to prenatal diagnostic testing (PDT) compared to not medically trained mothers. Moreover, we examine whether a change in guidelines affected the differences in care offered to expert and non-expert patients.
Methods: Our data is linked Danish administrative data on the use of PDT, patients age, gender, ethnicity education and family income from 51,204 mothers aged 33-37 giving birth from 1996 through 2002 and 23,211 mothers giving birth from 2008 through 2018. We use a differences-in-discontinuities design to estimate the difference in the use of pre-natal testing between expert and non-expert patients on the margin of a guideline threshold. We measure baseline preferences as the difference above the threshold, where all patients are offered PDT. Controlling for this baseline difference in preferences, we estimate expert “overuse” as the difference in the differences above and below the threshold. Prior to 2004 the threshold was age based (35 years) and after 2004 risk based (risk >1:300). We use exact matching to compare mothers with similar levels of education and income levels.
Results: We find a 7.4 percentage points overuse difference when the age-based threshold applied. Overall, 70 percent % of the difference in PDT is due to expert “overuse”. Experts and non-expert patients have similar test-rates above the threshold, indicating that the differences below the threshold are not driven by differences in preferences. After the risk-based threshold was introduced, the difference in PDT almost disappears.
Discussion: Expert mothers circumvent clinical guidelines intended to ration prenatal diagnostic testing indicating that the difference between experts and non-experts is due to agency discrimination.
25 March 2021, 14:00-15:00 (CET)
Objectives: Excessive length of hospital stay is one of the leading sources of inefficiency in healthcare. It can be caused by the lack of alternative care arrangements following a hospitalization. When a patient is medically fit to be discharged but requires some form of support outside the hospital (a short stay at a nursing home facility or home help), which is not readily available, the patient cannot be safely discharged. The patient stays at the hospital for a longer period until a safe discharge is possible –a phenomenon called bed-blocking. I investigate whether, and to what extent, the availability of publicly subsidized nursing homes (NH) and teams providing home care (HC) reduces hospital bed-blocking.
Methods: I use data on the universe of inpatient admissions at public hospitals in Portugal between 2000-2015. Portuguese public hospitals have no financial motivation to prolong lengths of stay. My main empirical analysis relies on a triple-differences design comparing the length of stay of individuals at increased risk of bed-blocking and the length of stay of patients not at risk of bed-blocking, before and after the entry of the first NH and HC team in their region of residence. This design exploits two distinct sources of variation. First, it exploits variation across regions and time in the availability of publicly subsidized NH and HC teams, originating from the staggered implementation of a policy reform. Second, it exploits variation between patients, who live in the same region and are admitted to the hospital in the same time period, in their propensity to bed-block, due to the presence of social needs (e.g. lack of family support).
Results: The entry of HC teams in a region reduces the length of stay of individuals at increased risk of bed-blocking by 4 days relative to regular patients –this reduces length of the bed-blocking period, but does not fully eliminate it. Reductions in length of stay upon the entry of NH occur only for patients with high care needs. Importantly, these reductions in bed-blocking do not come at a cost for patients’ health, showing that the bed-blocking period does not entail any meaningful care provision. Finally, the beds freed up by reducing bed-blocking do not remain unoccupied: there is an increase in the number of programmed admissions following the entry of HC teams, highlighting that longer waiting time for elective care are a relevant cost of bed-blocking.
Discussion: These results yield important policy implications. First, NH and HC teams target different patients and should be used as complements. Second, HC teams are more effective than NH at reducing bed-blocking because the average bed-blocker is not sick enough to need a NH. Taken together, my findings provide insights for organizing care delivery to patients with a complex combination of health and social needs.
8 April 2021, 14:00-15:00 (CET)
Public policy making for the prevention of diet-related disease is impeded by a lack of evidence on whether poor diets are a matter of personal responsibility or a choice set narrowed by environmental conditions. An important element of the environment are market imperfections in food retail, which may distort relative food prices and lead to suboptimal dietary choices. To identify such market imperfections, we exploit variation in diets across household that have different levels of income and live in different neighborhoods, using a rich dataset on quantities and prices of food purchases in the U.S. and a structural model of dietary choices. We find that distortions in prices are responsible for one third of the gap between the recommended and actual intake of fruit and vegetables. We construct a feasible fiscal intervention to remedy these distortions that makes all consumers better off.
15 April 2021, 14:00-15:00 (CET)
Objectives: A negative social gradient affects healthcare expenditures and longevity in opposite directions. Lower socioeconomic groups have higher current healthcare expenditures compared to higher socioeconomic groups, but also lower life expectancy. Since higher socioeconomic groups consume lower healthcare expenditures across a relatively long lifetime, it is unclear whether a negative health cost gradient exists in a lifetime perspective. This paper analyzes lifetime healthcare expenditures across socioeconomic groups using complete health cost data for all individuals in the entire Danish population.
Method: We calculate expected lifetime healthcare expenditures by gender and socioeconomic group, by weighting average healthcare expenditures by the percentage of the population alive to consume healthcare expenditures from age 30 to 100+.
Results: Contrary to existing literature, we find that all socioeconomic groups spend almost an equal amount on healthcare throughout a lifetime, once we account for socioeconomic mortality and healthcare expenditure differences simultaneously. On average, men in the lowest socioeconomic group spend $268,000 on healthcare across a lifetime, while men in the highest socioeconomic group spend $254,000. Meanwhile, females in the lowest and highest socioeconomic groups spend $352,000 and $328,000 respectively. Even though total lifetime healthcare expenditures across socioeconomic groups are near identical, the level of lifetime healthcare expenditures differs by cost component. The lowest socioeconomic group spends most on outpatient hospital care, prescription drugs, and about 20 percentage more on inpatient care compared to the highest socioeconomic group. Conversely, the highest socioeconomic group spends most on primary care physicians, home care plus home nurses, and 21 percentage more in nursing homes. Nonetheless, all socioeconomic differences in lifetime healthcare expenditures across any cost component are insignificant.
Discussion: Socioeconomic spending differences exist across the age dimension in average healthcare expenditures, but once we account for a full array of health costs and the longer average lifespan of those in the highest socioeconomic groups, the gradient vanishes. These findings question the effectiveness of healthcare systems with free universal healthcare as low socioeconomic groups have shorter life expectancies than higher groups even though they have almost equal lifetime healthcare expenditures. By cost component, socioeconomic groups have different average and lifetime spending, which characterize the health challenge facing each socioeconomic groups. The lowest socioeconomic group, for example, spends most on in- and outpatient hospital care, while the highest socioeconomic group spends most on elderly care. We suggest that policymakers leverage these differences to create more effective policies.
22 April 2021, 14:00-15:00 (CET)
This paper analyzes the long-term effects on mortality and socio-economic outcomes from being born in a maternity ward compared to home births. We focus on two Swedish interventions that affected the costs of hospital deliveries and the supply of maternity wards during the 1926–46 period. Using exogenous variation from the supply of maternity wards to instrument the likelihood of giving birth in an institution, we find that giving birth in a maternity ward has substantial effects on later-life outcomes such as educational attainment and mortality.
We argue that a decrease in child morbidity from better treatment in the case of complications could be a likely explanation for the large gains from being born in a hospital. This interpretation is corroborated by evidence from primary school performance, which show a large reduction in the probability of performing in the left tail. In contrast to an immediate and large take-up in hospital deliveries as response to an increase in the supply of maternity wards, we find no increase in hospital births from the abolishment of fees – but instead some degree of displacement of high-SES parents.
29 April 2021, 14:00-15:00 (CET)
This study examines the impact that pharmaceutical innovation, which accounts for most private biomedical research expenditure, has had on longevity. We perform two types of two-way fixed-effects analyses, which control for the effects of many potentially confounding variables.
First, we analyze long-run (2006-2018) changes in longevity associated with different diseases in a single country: the U.S. Then, we analyze relative longevity levels associated with different diseases in 39 countries during a single time period (2006-2016). The measure of longevity we analyze, mean age at time of death, is strongly positively correlated across countries with life expectancy at birth. The measure of pharmaceutical innovation we use is the mean vintage (year of initial world launch) of the drugs used to treat each disease in each country. Changes in the vintage distribution of drugs are due to both entry of new drugs and exit of old drugs.
Our analysis of U.S. data indicates that the diseases for which there were larger increases in drug vintage tended to have larger increases in longevity of all Americans and of Black Americans. In other words, the lower the mean age of the drugs, the higher the mean age at death. We test, and are unable to reject, the “parallel trends” hypothesis. The estimated effect of drug vintage on Black longevity is 26% larger than the estimated effect of drug vintage on overall longevity. We estimate that the 2006-2018 increase in drug vintage increased the mean age at death of all races by 0.79 years (103% of the observed increase), and that it increased the mean age at death of Black Americans by 0.88 years (54% of the observed increase).
Our analysis of data on 39 countries indicates that the higher the vintage of drugs available to treat a disease in a country, the higher mean age at death was, controlling for fixed disease and country effects. The increase in drug vintage is estimated to have increased mean age at death in the 39 countries by 1.50 years between 2006 and 2016. This increase is almost twice as large as our maximum estimate of the 2006-2018 increase in mean age at death in the U.S. attributable to the increase in drug vintage (0.79 years). However, longevity increased much more throughout the world than it did in the U.S., so the increase in drug vintage accounted for 36% of the overall increase in mean age at death in 39 countries.
Controlling for the mean vintage of the drugs sold in a country during 2008-2016, the larger the fraction of those drugs that were first launched in that country after the year 2006, the lower longevity was, presumably because utilization of a drug tends to be quite low during the first few years after it first becomes available, and also because drugs may have to be consumed for several years for their full health benefits to be realized.
6 May 2021, 14:00-15:00 (CET)
The concentration of healthcare spending at the end of life is widely documented but poorly understood. To gain insight, we focus on patients newly diagnosed with cancer. They display the familiar pattern: even among cancer patients with similar initial prognoses, monthly spending in the year post diagnosis is over twice as high for those who die within the year than those who survive. This elevated spending on decedents is almost entirely driven by higher inpatient spending, particularly low-intensity admissions, which rise as the prognosis deteriorates. However, even for patients with very poor prognoses at the time of admission, most low-intensity admissions do not result in death, making it difficult to target spending reductions. We also find that among patients with the same cancer type and initial prognosis, end-of-life spending is substantially more elevated for younger patients compared to older patients, suggesting that treatment decisions are not exclusively present-focused. Taken together, these results provide a richer understanding of the sources of high end-of-life spending, without revealing any natural “remedies.”
Speaker: Dan Zeltzer, Tel Aviv University
Discussant: Eric French, University of Cambridge
Chair: Luigi Siciliani, University of York
20 May 2021, 14:00-15:00 (CET)
Objectives: As in other European countries, users of residential care in Austria were required to contribute to the costs of residential care based on both their income and assets. This asset-based out-of-pocket payment (OPP) for residential care – denominated Pflegeregress – was abolished in 2018. There is currently no empirical evidence on the distributional effect of this measure across income groups. This study thus aims to answer the following research questions: how was Pflegeregress distributed across different income and home ownership groups before 2018 and what were the distributional consequences of its abolishment?
Methods: The study uses a matched administrative (residential care users and mortality tables) and survey (SHARE, wave 6 for Austria) dataset on which a purposely built micro-simulation model was applied. The micro-simulation model estimates the annual OPPs for residential care borne by individuals 65+ for a reference year (2015). The distributional impact is assessed through a series of measures, including Concentration Indices (CI) and Curves (CC) and OPPs adjusted for need.
Results: The findings from the micro-simulation model show that in absolute value, users from the upper income quintile paid the highest total OPPs. However, asset-based OPPs (i.e. Pflegeregress) represented both a higher share of total OPPs paid and a higher proportion of the financial resources of lower income users, with a large share of the Pflegeregress financed through housing assets of home owners. The CC show that the 20% poorest individuals accounted for 25% of the estimated Pflegeregress paid in 2015. This is in contrast with the majority of higher-income individuals who could cover residential care fees from their income alone, without resorting to assets. Adjusting for need shows that this distribution was very much driven by the concentration of residential care in less affluent individuals. The abolishment of the asset-based OPPs for residential care in Austria thus benefited mostly those in the lower income quintiles.
Discussion: The Pflegeregress fell disproportionately on low-income residential care users, which means that the regressive nature of Pflegeregress made this group the largest beneficiary of its abolishment. Given how the distribution of asset-based OPPs was influenced by residential care use and duration – both concentrated on less affluent individuals – it is questionable whether asset-based OPPs in residential care are able to target payments to wealthier individuals. Other financing options such as earmarked inheritance taxes or social insurance schemes that decouple financing from use could be alternatives to circumvent the adverse distributional effects of asset-based OPPs. Our findings show the relevance of considering the distributional implications of different financing options for care.
27 May 2021, 14:00-15:00 (CET)
How is the prenatal sex selective behaviour influenced by the presence of cheap fetal gender identification technology and financial incentives? We study this question by analysing a conditional cash transfer program called Janani Suraksha Yojna (JSY) implemented in India. By providing access to prenatal sex detection technology like the ultrasound scans and simultaneously providing cash incentives to both households and community health workers for every live birth, this program altered existing trends in prenatal sex selection. Using the difference-in-difference and the triple difference estimator we find that JSY led to an increase in female births by 4.8 and 12.7 percentage points respectively. Additionally, the likelihood of under 5 mortality for girls born at a higher birth order increased by around 6 percentage points. Our calculations show that this resulted in nearly 300,000 more girls surviving in treatment households between 2006 and 2015. We find that the role played by community health workers in facilitating this program is a key driver of decreasing the prenatal sex selection.
3 June 2021, 14:00-15:00 (CET)
In the healthcare sector, the outcome-volume relation can be explained by two dynamics: the learning by doing and the selective referral hypotheses. The former implies the more a task is performed the better the outcomes achieved, while the latter suggests hospitals with better performances attract more patients, thus leading to higher volumes of treatments. This study aims to investigate the relation between the volume of elective surgeries and health outcomes for patients undergoing hip replacement surgery. Elective surgeries are operations that are scheduled in advance. Hence, patients can select in which hospital to be treated, making the hospital choice nonrandom. This work attempts to disentangle the two contrasting mechanisms. To address patient nonrandom selection, a conditional choice model is implemented where the probability of hospital selection is estimated based on patient and hospital characteristics, and travel distance from residency to hospital. Patients are expected to choose the closest hospital only if the unobserved quality is homogeneous in the region. Thus, traveling longer distances signals higher perceived hospital quality. The estimated probability of choosing each hospital is used to predict hospital volume, purged of the selective referral component, which is then included in the health outcome equation. Findings suggest the presence of both learning by doing and selective referral. Policies aiming to aggregate provision may be effective, but caution should be posed to the dynamics of these two contrasting mechanisms.
10 June 2021, 14:00-15:00 (CET)
As a consequence of the Spring 2020 lockdown that occurred in Spain due to the COVID-19 pandemic, many people lost their jobs or had to be furloughed. The objective of this study is to analyse the influence of the latter changes in labour market status on psychological well-being. For this purpose, an ad-hoc questionnaire featuring socio-demographic and mental health criteria was created. Granted that the pandemic can be viewed as an exogenous shock, the bias caused by the bidirectional problems between the work situation and mental well-being can be tackled. Results indicate that the lockdown exerted a greater negative effect on the self-perceived well-being of unemployed and furloughed persons than on those in employment. Moreover, among those in continuous employment, teleworkers experienced a lesser degree of self-perceived well-being post lockdown as compared to those people remaining in the same work location throughout the COVID-19 crisis. Finally, the lockdown provoked worse effects on the self-perceived well-being of women as compared to men, a result that appears to be related to gender differences in household production. In conclusion, these results could be especially relevant given that the evolution of the pandemic is having ongoing effects on employment and, therefore, on the mental health of workers.
17 June 2021, 14:00-15:00 (CET)
This paper evaluates the effects of the universal poverty database (UDB) reform in Indonesia in 2013 - intended to improve poverty targeting and to extend health insurance coverage for the poor. The reform aimed to identify the 40 percent poorest households. We employ an event study method to estimate the impacts of the reform on health insurance coverage across “poverty” quintiles, measured by per capita consumption. We then also examine the effects of the reform on the healthcare utilisation and private spending on health.
Our results suggest that the reform i) did indeed increased insurance coverage among the poorest quintile substantially - 14 percentage points in the first quarter of 2013 and 17 percentage points in the last quarter of the year; ii) raised health insurance coverage among the second poorest quintile by around 10 percentage points immediately and 14 percentage points in the following quarters. Yet, while successful in targeting the poor, the estimated effects of the reform on subsequent quintiles point to substantive “leakage” in targeting: health insurance coverage through Jamkesmas was intended to provide free insurance only to the poorest 40 percent. Yet, health insurance after the reform also increased across the richer household groups, albeit to a lesser extent and impacts are linearly decreasing across quintiles.
Despite improvement in the health insurance rates, we find limited improvement on the healthcare utilisation after the reform across household quintiles. This result stays unchanged after examining effects of the reform on healthcare utilisation between districts with high and low supply of medical providers. It shows that current variations in healthcare infrastructure cannot explain the limited improvement in the healthcare utilisation services. Consistent with our finding on health insurance rates, households’ out-of-pocket spending on health decreased for approximately 20 percentage points suggesting that the reform reduced the financial burdens on households.