Working Papers
Measuring Beliefs and Uncertainty about Medical Benefits and Costs in Health Care Decisions (with Andreas Landmann and Lisa Rogge) - (submitted) (earlier version available as BGPE Discussion Paper No. 236)
Abstract: In a world with incomplete financial health protection, uncertainty about medical outcomes as well as about costs to seek care might play an important role in the health care decision-making process, potentially deterring sensible health care choices. There is no rigorous empirical evidence on this relationship though, also owing to the lack of established tools to measure uncertainties around both medical benefits and costs of health care seeking. In this paper, we introduce a new ‘double uncertainty’ measurement tool based on health vignettes and present results from fielding it in a low-income population in Pakistan. This first empirical data reveals that on top of biases, both medical as well as cost uncertainty are present in the target population in health scenarios where seeking professional care is medically necessary. We qualitatively and quantitatively validate the new measurement tool, showing that it is both valid and reliable. We furthermore show that our measures can help to explain health care decisions, which indicates its usefulness for future applications.
Selected Work in Progress
Costs and Benefits of Health Care Decisions: Biases, Uncertainty, and Information (with Andreas Landmann, Richard Peters, and Lisa Rogge) - preregistered in AEA registry
Abstract: Health care decisions are often made under incomplete information about medical benefits and financial consequences. Such frictions are especially important for low-income households, for whom delayed care can be life-threatening, while costs of care might be financially devastating. Based on predictions from a simple theoretical framework, we study biases and uncertainty using a validated survey instrument that elicits beliefs about treatment benefits, expected costs, and perceived uncertainty for hypothetical health shocks among 3,397 low-income households in Pakistan. We document four main findings. First, misperceptions about benefits and costs are widespread: respondents overestimate the effectiveness of lower-level care for severe conditions and expect sizable costs even for services that should largely be covered by public inpatient insurance. Second, treatment decisions are often inappropriate, with only 31% of respondents making the fully appropriate choice for symptoms of a heart attack and 17% for appendicitis. Third, perceived benefits and expected costs predict provider choice in the expected direction, especially for severe conditions, whereas measured uncertainty plays only a limited role despite its theoretical relevance. Fourth, a targeted information intervention reduces some cost misperceptions at the margin, but has little effect on medical beliefs or high-stakes treatment decisions. Our findings suggest that informational frictions in this setting are substantial, but unlikely to be resolved by one-shot information provision, even when the information is highly targeted.
Health Insurance Preferences for Outpatient Care: A Discrete Choice Experiment in Pakistan (with Henrik Berberich and Lisa Rogge) - (submitted, working paper available upon request)
Abstract: State-funded health insurance schemes are increasingly implemented across low- and middle-income countries, but utilization and acceptance often remain lower than desired for Universal Health Coverage. Including features that address the beneficiary population’s preferences could improve this. We conducted a Discrete Choice Experiment to elicit preferences for a new public outpatient health insurance for low-income households in Pakistan at scheme design stage. We included five attributes that reflected the dimensions of real policy trade-offs during scheme design: health care providers, services, health conditions, coverage amount and premium. The main effects reveal relevance of all attributes and strong population preferences for including higher-level health care providers as well as telemedicine and for covering chronic disease needs. We do not detect substantial heterogeneity in preferences across socio-demographic strata, respondent and household health status, but some regarding gender and location. This indicates that overall, preferences are rather homogenous across groups, but preferences of females and respondents from more remote areas differ. We further see suggestive evidence that in a setting with low insurance literacy, only some DCE choices were made to maximise benefits along known, pre-existing health complaints and risk-factors.
Published work
Un/met: A mixed-methods study on primary healthcare needs of the poorest population in Khyber-Pakhtunkhwa province, Pakistan (with Maira Shaukat, Lisa Rogge, Fatima Khalid, Safat Ullah, Fayaz Ahmad, Zeeshan Kibria, Andreas Landmann, Zohaib Khan, Manuela De Allegri) - International Journal for Equity in Health (2024)
Reports
Needs Assessment for Scientific Implementation Research on OPD Services in the Social Health Protection Initiative in Khyber Pakhtunkhwa (KP) Province and Gilgit Baltistan (GB) Area (Baseline Report for German Development Bank) University of Erlangen-Nuremberg