Module 1: Hospital Pricing and Selection on Observables

class: center, middle, inverse, title-slide

# Module 1: Hospital Pricing and Selection on Observables
## Part 1: HCRIS Data
### Ian McCarthy | Emory University
### Econ 470 & HLTH 470

---

<style type="text/css">
.remark-slide-content {
    font-size: 30px;
    padding: 1em 2em 1em 2em;    
}
.remark-code {
  font-size: 15px;
}
.remark-inline-code { 
    font-size: 20px;
}
</style>

# What is a hospital price?

Defining characteristic of hospital services: *it's complicated!*

--
.center[
  ![:scale 800px](pics/BillExample.jpg)
]

<div class="smalltext">Brill, Steven. 2013. "Bitter Pill: Why Medical Bills are Killing Us." *Time Magazine*.</div>

---
# What is a hospital price?

Lots of different payers paying lots of different prices:
- [Medicare fee-for-service prices](https://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNProducts/Downloads/AcutePaymtSysfctsht.pdf)
- [Medicaid payments](https://www.kff.org/report-section/understanding-medicaid-hospital-payments-and-the-impact-of-recent-policy-changes-issue-brief/)
- Private insurance negotiations (including Medicare Advantage)
- But what about the price to patients?

.center[
Price `\(\neq\)` charge `\(\neq\)` cost `\(\neq\)` patient out-of-pocket spending
]

---
# What is a hospital price?

.center[
  ![:scale 600px](pics/DifferentPrices.jpg)
]

<div class="smalltext">Source: <a href="https://healthcarepricingproject.org/">Health Care Pricing Project</a></div>

---
# What is a hospital price?
Not clear what exactly is negotiated...

--
.pull-left[
### Fee-for-service
- price per procedure
- percentage of charges
- markup over Medicare rates
]

--
.pull-right[
### Capitation
- payment per patient
- pay-for-performance
- shared savings
]

---
# Hospital prices in real life
A few empirical facts:

1. Hospital services are expensive

2. Prices vary dramatically across different areas

3. Lack of competition is a major reason for high prices

---
# Hospital prices in real life

.pull-left[
  ![:scale 450px](pics/HC_var_withinmkt_hip_ga_atlanta.png)
]

.pull-right[
  ![:scale 450px](pics/HC_var_withinmkt_kmri_ga_atlanta.png)
]

<div class="smalltext">Source: <a href="https://healthcarepricingproject.org/">Health Care Pricing Project</a></div>

---
class: inverse, center, middle
name: hcris

# Understanding HCRIS Data

---
# What is HCRIS?
Healthcare Cost Report Information System ('cost reports')
- Nursing Homes (SNFs)
- Hospice
- Home Health Agencies
- Hospitals

---
# Hospital Cost Reports

.center[
  ![:scale 800px](pics/HCRIS.png)
]

---
# The Data

Let's work with the [HCRIS GitHub repository](https://github.com/imccart/HCRIS). But forming the dataset is up to you this time.

--
.center[
  ![:scale 700px](https://media.giphy.com/media/26DNdV3b6dqn1jzR6/giphy.gif)
]

---
# The Data

```r
hcris.data %>% 
  ggplot(aes(x=as.factor(year))) + 
  geom_bar() +
  labs(
    x="Year",
    y="Number of Hospitals",
    title="Number of Hospitals per Year"
  ) + theme_bw() +
  theme(axis.text.x = element_text(angle = 90, hjust=1))
```
.plot-callout[
<img src="01-hospital-pricing1_files/figure-html/hospital-count-callout-1.png" style="display: block; margin: auto;" />
]

---
# Number of hospitals

---
# Estimating hospital prices

```r
hcris.data <- hcris.data %>%
  mutate( discount_factor = 1-tot_discounts/tot_charges,
          price_num = (ip_charges + icu_charges + ancillary_charges)*discount_factor - tot_mcare_payment,
          price_denom = tot_discharges - mcare_discharges,
          price = price_num/price_denom)
```

---
# Estimating hospital prices

.left-code[

```r
hcris.data %>% group_by(year) %>% 
  filter(price_denom>10, !is.na(price_denom), 
         price_num>0, !is.na(price_num)) %>%  
  select(price, year) %>% 
  summarize(mean_price=mean(price, na.rm=TRUE)) %>%
  ggplot(aes(x=as.factor(year), y=mean_price)) + 
  geom_line(aes(group=1)) +
  labs(
    x="Year",
    y="Average Hospital Price",
    title="Hospital Prices per Year"
  ) + scale_y_continuous(labels=comma) +
  theme_bw() + theme(axis.text.x = element_text(angle = 90, hjust=1))
```
]

.right-plot[
![](01-hospital-pricing1_files/figure-html/price-plot1-1.png)
]

---
# Estimating hospital prices

.left-code[

```r
hcris.data %>% group_by(year) %>% 
  filter(price_denom>100, !is.na(price_denom), 
         price_num>0, !is.na(price_num),
*        price<100000) %>%
  select(price, year) %>% 
  summarize(mean_price=mean(price, na.rm=TRUE)) %>%
  ggplot(aes(x=as.factor(year), y=mean_price)) + 
  geom_line(aes(group=1)) +
  labs(
    x="Year",
    y="Average Hospital Price",
    title="Hospital Prices per Year"
  ) + scale_y_continuous(labels=comma) +
  theme_bw() + theme(axis.text.x = element_text(angle = 90, hjust=1))
```
]

.right-plot[
![](01-hospital-pricing1_files/figure-html/price-plot2-1.png)
]