Blog

Weekly Preliminary OS Climate Risk Review (10/7/21)

By October 14, 2021 No Comments

This week we take a closer look at Sentara Healthcare which has a concentration of hospital campuses sitting in the Hampton Roads region of Virginia, a pair of campuses located further inland west of Richmond, as well as a medical center with 100th percentile flood exposure straddling the coast of North Carolina. We dig into the underlying socioeconomic profiles within the service areas of the hospitals with higher climate risk, and unearth recent allegations of fraud and HIPAA violations brought against the System. Spoiler alert: you’re not going to find any discussion in this week’s preliminary official statement pertaining to environmental risk, socioeconomic vulnerabilities or past episodes of administrative malfeasance.

Let’s get to work, pass the scalpel please. 

This week’s non-disclosure patient on the operating table:

  • Sentara Healthcare, Taxable Bonds Series 2021 ($300,000,000); 
  • Economic Development Authority of Rockingham County, Virginia (Sentara Healthcare) ($145,855,000) Series 2021A

 The Bonds (Sentara Portfolio risQ Score 1.7; Flood risQ Score 3.4; Hurricane risQ Score 1.6)

The Bonds are general obligations of Sentara Healthcare with the proceeds being used for general corporate purposes, facility equipment, and to refinance previously issued debt. The Obligated Group consists of 8 hospitals from Sentara Hospitals, as well as Sentara RMH Medical Center, Martha Jefferson Hospital and Northern Virginia Medical Center. Sentara’s Albemarle Medical Center is the only hospital site located outside of Virginia, and sits on the coast of North Carolina. The series is available through a $300 million offering and as an approximately $146 million offering through the Economic Development Authority of Rockingham County; the System’s debt service requirements are currently dated to 2051 (preliminarily). 

Non-disclosure

The BONDHOLDERS’ RISKS section in Sentara’s POS spans from page 18 to 49, and covers myriad topics including COVID-19, cybersecurity and reliance on medicare. On page 28, squished between a section on the risks posed by labor costs and infectious disease outbreaks, we find the a shallow two-sentence weather risk disclosure section:

Facility Damage. Health care facilities are highly dependent on the condition and functionality of their physical facilities. Damage from natural causes, severe weather, fire, deliberate acts of destruction, terrorism or various facility system failures may have a material adverse impact on the System’s business and financial condition, especially if insurance is inadequate to cover resulting property and business losses.

We couldn’t help ourselves: copying the text above and pasting it into Google yields a handful of results for previously published official statements that use nearly identical disclosure language, despite having entirely different climate risk profiles:

Beyond the copy-and-paste section, there is no mention of flood risk or climate change, despite the fact that 7 of the 11 hospital sites have materially high coastal flood exposure. The only mention of “flood” is buried in the force majeure clause in Appendix C.

Hurricane and Coastal Flood Exposure

Unlike some hospital systems that have campuses spread across the country, it’s worth noting that this system is regional. This means that a serious hurricane or tropical storm event creates risk that could be correlated among the hospitals and surrounding populations. Though the Hurricane risQ Score does not necessarily stand out at first glance, it is worth recalling 2018’s Hurricane Florence and the havoc it wreaked on the Carolinas. The storm gained infamy for stalling and dropping massive amounts of rainfall inland and causing ~$24B in damages. And, the recovery from Florence is still far from over. The Sentara System is slightly north of where Florence traveled, and as such, hurricane landfall likelihood is only ~half as likely. Many of the System’s hospitals are close to the coast and as such are exposed to the triple threat of wind, rain, and storm surge, in the event of a Florence-like storm, and most of the campuses and their service areas are exposed to significant hurricane flood damage a Florence-like storm were to occur. If there was any further doubt, this year’s Ida has shown that we should no longer be thinking about hurricanes as only a coastal threat. And of course, beyond explicit hurricanes, more moderate storm systems and coastal flooding pose a serious and consistent threat to the area; Albemarle in North Carolina stands out as among the highest risk in the US.

Coastal flooding is the largest climate risk factor for the health system. At the 30-minute drive time radius, 6 of the System’s 11 hospitals rank >=90th percentile for property losses from coastal flooding, and 4 hospitals rank >=90th percentile for GDP impairment. Albemarle Medical Center is located near the inlets of the Albemarle Sound and ranks in the 100th percentile nationally for property losses resulting from combined climate perils, with a 100-year coastal flood event expected to drive losses of 27% for the area within a 30-minute drive of the campus. At Sentara’s Virginia Beach, Princess Anne, Leigh, and CarePlex locations, a 100-year coastal flood is projected to drive losses of 6-7% within the 30-minute band.

Flood Risk, Housing Affordability and Health Challenges

It’s critical to consider both flood exposure as well as Social Impact factors in order to obtain a complete view of Sentara Healthcare’s aggregate climate risk. Hospitals partially rely on discretionary income to drive revenue, which is both diverted towards property repairs and adversely impacted by the shocks to local economies following a climate event; decreases in discretionary income translate to decreases in elective health procedure revenues for hospitals, meaning potentially less revenue for Sentara, and potentially follow-on credit issues for the System in future years. Let’s focus in on the population health and housing affordability profiles for the most flood-prone hospitals in the Sentara Healthcare system (bed counts and 30-minute drive time band Flood risQ Scores provided below):

  • Sentara Albemarle Medical Center (182 beds, 4.9)
  • Sentara Leigh Hospital (247 beds, 3.9)
  • Sentara Norfolk General Hospital (525 beds, 3.7)
  • Sentara Princess Anne Hospital (174 beds, 3.9)
  • Sentara Virginia Beach General Hospital (273 beds, 4.1)

Percent of Income Spent on Housing Score (acronym PISH)— higher Scores are seen in areas with populations that on average spend a higher percentage of income on housing. Below we provide a PISH view of the 5 highest flood-prone Sentara campuses, all of which have community’s suffering from affordable housing issues within the 6-minute drive time band, Scoring at least above 80. (*Given the fact that all hospitals Score above 75 across all drive time bands, we’re narrowing the Score range on the vertical axis to 75100). As we slide that drive time radius out to 90-minutes, the 5 hospitals (which are all located in relatively close geographic proximity to one another) converge at a Percent of Income Spent on Housing Score of ~83.

The 525-bed Norfolk General Hospital (Flood risQ Score 3.7) has a Percent of Income Spent on Housing Score of 93 within a 6-minute drive time of the campus, showing that the cost of living within walking distance of the academic hospital is already too high. Norfolk General’s PISH in fact increases to 96 at a 9-minute radius, and then settles back down around 87 at the 30-minute radius — within that 30-minute band, Average Owner occupied housing unit Real Estate Taxes ranks 94th percentile and Median Monthly Housing Costs ranks 78th percentile, nationally. If coastal flooding events continue to increase in severity and frequency as result of warming sea surface temperatures and heavier inland flood, the community surrounding Sentara’s largest hospital (by bed count) could see the cost of living reach unsustainable levels. These affordability issues are expected to worsen as insurance costs increase as a function of FEMA’s transition to NFIP Risk Rating 2.0 (“RR2.0”), which is calibrating insurance premiums to actual flood risk and is expected to drive up premiums in Special Flood Hazard Areas (SFHAs). Given the fact that NFIP rate increases are capped at 18% per year, it is expected that approximately 50% of policyholders within the region will see rate increases of only less than $10 per month, according to the Hampton Roads Planning District Commission. However, for communities with the highest flood risk — such as those living near the basket of 5 Sentara hospitals — it is expected to take 5-10 years for most policies to reach their loss-based premium rates, and even as long as two decades for those properties with the largest premium increases, according to FEMA.

The inflationary impacts of increasing insurance rates driven by worsening floods might not hit homeowners next year; but, there will undoubtedly be a slow burn of household budgets and discretionary spending that could have adverse consequences for the region over the next 10-20 years.

Persistent Health Obstacles Score (acronym PHO) — higher scores are seen in areas where the populations have high rates of underlying health conditions/risks and in some cases less access to health insurance. Below we provide a view of the PHO Scores for each of our 5 focus hospitals, across the entire drive time band range from 6- to 90-minutes. (*The Score range on the vertical axis is provided from 0-100).

Critical insights drop out when we compare our Social Impact Scores representing both population health and housing affordability, as well as other key socioeconomic indicators, for the communities residing within the Sentara service areas:

  • Sentara Albemarle Medical Center (182 beds, 4.9): the only hospital in the System located in North Carolina, along the Pasquotank River in Elizabeth City. Both its PISH and PHO Scores sit around 90 within the 6-min band, and gradually decrease to below 80 within 30-min. In the event of a catastrophic flooding event, Albemarle Medical Center (whose service area ranks 100th percentile for total property losses within the healthcare sector) has the most vulnerable population from a socioeconomic standpoint. Within the 20-min drive time band, Albemarle’s PISH Score is 82 and PHO Score is 80, indicative of a population in need of preventative health services, although may have difficulty accessing or affording such services in the event of a hurricane or flood that drives up the cost of living and erodes discretionary income levels within the area.
  • Sentara Leigh Hospital (247 beds, 3.9): located in Norfolk VA, its PHO Score drops and remains below 20 beyond the 25-min drive time band, an indication that the healthier populations live further from the hospital. Despite a PISH Score of 87, Leigh’s socioeconomic profile appears more robust than Albemarle’s within the 25-min drive time band, with a Low Prestige Employment Score of 20 and Low Affluence Score at 31 (Albemarle Scores are 46 and 59, respectively). In the event of a major climate event, the population living within Leigh’s surrounding area appears to have a higher level of economic resilience than Albemarle’s and may be better situated to bounce back.
  • Sentara Norfolk General Hospital (525 beds, 3.7): the largest hospital by bed count in the Sentara System, located on the Elizabeth River about a 20-minute drive from Leigh Hospital. Norfolk General’s PISH Score is a polar opposite to its PHO Score within a 6-min drive time band (93 and 6.6, respectively), showing that the community living closest to the major hospital primarily comprises a healthier population, most likely either attending or teaching at the nearby Eastern Virginia Medical School (33% of jobs are in the Health Care and Social Assistance sector, ranking 85th percentile, and 16% of jobs are in Educational Services, ranking 86th). As we expand out to the 12- and 15-minute drive time bands, PHO pops from single digits to as high as 89, showing that the community within reasonable driving distance of Norfolk General has a heightened need for health services, relatively high lack of health insurance (17% for the 18-64 year age group, ranking 63rd percentile), as well as a high cost of living. We also see high Poverty Concentration and Nonwhite/Minority Population Scores of 62.2 and 92.3 for the population within the 15-minute drive time band. Norfolk General appears to have the highest bed count for good reason.
  • Sentara Princess Anne Hospital (174 beds, 3.9): named after Anne Stuart, Queen of Great Britain during the first decade of the 18th century and namesake to the encompassing Princess Anne County, the Hospital has some landmass insulation from the Elizabeth River and Atlantic Coast. The Hospital’s POH Score starts low at 4.3 within the 6-min, and gradually increases to 5.5 at the 15-min and 16.0 at the 30-min, before converging with the other hospitals in the mid-30s at the 90-min. Unlike Norfolk General above — which has a population with noteworthy health vulnerabilities living within the service area from a 9- to 20-min drive from its campus — the population within the driving distance of Princess Anne is generally healthier given its younger more affluent profile (19% of population is considered high income young homeowners, ranking 87th percentile nationally).
  • Sentara Virginia Beach General Hospital (273 beds, 4.1): sitting on the Linkhorn Bay near the Atlantic Coast, Virginia Beach General has a similar POH profile as Princess Anne above, which is located ~25-min drive southwest. Virginia Beach General’s POH Score lingers in the single-digits between the 6- to 20-min bands. That population within the 20-min band is relatively affluent (only 9% below poverty line, ranking 11th percentile nationally), younger in age (26% of jobs for workers <29 years old, 90th percentile nationally), many working in sectors adversely impacted by weather events (14% work in Accommodation and Food Services, ranking 90th percentile nationally). Given that the population is relatively young, wealthy and mobile, one potential concern for Virginia Beach General is that increasing housing costs could drive its higher-dollar patient population from the area if flood risk worsens.

Governance concerns

Rounding out the ESG holy trinity with a brief mention of Sentara’s poor disclosure of administrative wrongdoing, we provide a handful of historical instances of fraud, HIPAA violations, employee malfeasance, and CMP law violations at Sentara:

Now, Sentara isn’t special — fraud is prevalent in the healthcare industry, and there are offenders all over the place (for instance, Indiana University Health cited above was forced to pay $18 million to resolve allegations that the hospital was involved in an illegal kickback scheme in 2017).  It’s not the fraud we’re surprised by, but rather the blatant lack of disclosure that’s so offensive. Although Sentara self-disclosed its most recently investigated misconduct to the HHS’ Office of the Inspector General, there is no mention of that offense in Sentara’s preliminary official statement. There are boilerplate sections describing Government Fraud Enforcement (pg. 26), the False Claims Act and Federal Fraud and Abuse Provisions (pg. 39), but not a single mention of the Civil Monetary Penalties (CMP) Law that Sentara violated in July (yes, this July).

While evaluating the Sentara series, one should factor in the lack of disclosure of the System’s exposure to flood risk, how those risks could affect patients’ discretionary income and hospital revenues, past government misconduct, as well as the socioeconomic profiles of the areas that the hospitals serve — sometimes Thursday non-disclosure surgery can surface insights on an ESG offender that couldn’t overwise be gleaned from its sickly preliminary official statement.

Honorable Mentions

Fontana CFD No. 31 (Citrus Heights North), CA (risQ Score 3.4, Wildfire risQ Score 4.2) $8,020,000

Following the copy-and-paste trend of California issues that we highlighted two weeks ago (Eagle Crest and Menifees), the POS begins with a generic statement of wildfire risk in California (pg. 52). However, the POS does provide some geographic specificity, mentioning major wildfires in the vicinity of the city over the past 20 years (pg. 52). But, the POS needs to go farther; this assessment fails to disclose nuances in geospatial wildfire risk between the city and the CFD. Fontana CFD No. 31 sits north of the city close to wildland vegetation meaning that compared to the city, the CFD is riskier because it’s nearer to the wildland-urban interface. The CFD has a Wildfire risQ Score of 4.2 compared to just 2.4 in the city. A 100-yr wildfire in the CFD drives 30% of losses and a 500-yr wildfire drives upwards of 73% of losses. These percentages drop to 5% and 11% respectively for the City of Fontana. The Grand Prix Fire in 2003, which the POS mentions (pg. 52), spared the city of any damage, but the burn perimeter was less than a mile from the CFD. Increased geographical focus is needed to disclose wildfire risk. The CFD also faces considerable drought (pg. 51) and heat stress. By 2050 the number of days expected to be above 95˚F increases by 50-75 days under RCP 8.5 conditions. In the same timeframe and climate scenario, the percent of months in extreme drought is expected to rank in the 95th percentile nationally. 

Once again, we see the intersection of housing, race, and wildfire risk. The CFD has a Minority/Nonwhite Score of 86.4 and a Percent of Income Spent on Housing Score of 82.3. Minority communities and housing-cost burdened communities are disproportionately affected by wildfire risk. Portions of the CFD are currently under development to provide for an additional 1,000 residential units. Though municipalities are facing a housing shortage, development in the wildland-urban interface seems ill-advised and makes more communities vulnerable.

Brazoria County, TX (risQ Score 3.8, Flood risQ Score 3.8, Hurricane risQ Score 3.1) $91,070,000

You would think that, with $91 million in GO bonds going toward renovating/repairing County facilities, the issuer would have a proactive stance on climate until the bonds’ maturation year in 2051. They are transparent about multiple 500-year flood events since 2015, including damages to County facilities from Harvey in 2017, so why not address mitigation efforts? Without a plan, the POS reads like they’re going to flush $91 million away. 

Brazoria County is located in the Gulf Coast region of Southeast Texas called the Texas Coastal Bend. The area is notorious for hurricanes and the torrential rain and storm surge that accompanies them. Last month, residents of Brazoria were hit hard by Hurricane Nicholas, with a reported 460,000 residents affected by power outages. If physical risk posed the singular threat to the County, it would require an immediate proactive strategy. Unfortunately for Brazoria, physical risk is accompanied by Carbon Transition Risk, drought, and lackluster disclosure from the issuer. The 154-page POS manages to cram the discussion of climate change risk—as it relates to bonds that mature over the course of three decades—into a few paragraphs that are shorter than this piece (and less informative) on pages 23-24. In fact, this is the only section that features the words hurricane and storm. Similar to all boilerplate mentions on climate, they acknowledge the potential threats, yet completely ignore what should be done to mitigate damages. 

Some might wonder how much damage a 500-year event causes. According to our projections, a 500-year inland flood event drives 15% property losses. Brazoria County is ranked 98th percentile for property losses from inland flooding, and 97th percentile nationally for combined physical hazards. In 2051, the County will have a Total risQ Score of 4.5. Severe and extreme drought ranks in the upper 90th and 80th percentiles, respectively, out until 2050 as well. Brazoria’s top ten taxpayers are chemical, petroleum, and industrial manufacturers (pg 14). These companies contribute to the County’s total scope 1 per capita emissions and ranks 92nd percentile nationally. Transition away from these energy sources will likely see a disruption in the relatively low Social Impact Score of 25.9 as people seek out alternative employment and living opportunities. With these metrics disclosed, adopting and disclosing a mitigation/adaptation strategy seems paramount.

CommCare, LA (risQ Score 2.3, Flood risQ Score 2.6) $184,895,000

CommCare is obligating 11 CCRC spread throughout LA for this bond series aimed at acquiring 4 additional facilities (which are included in the obligated group). While the geographic diversity protects CommCare from a single climate-event impairing all their facilities, recent events have been damaging to some. The POS is forthcoming with the laundry list of hurricanes that have made landfall in LA in recent years (pg. 37); this year Hurricane Ida made its presence felt by damaging two of CommCare’s facilities, while three more experienced power outages from the storm (pg. A-27).

5 facilities are above the 87th percentile nationally for GDP impairment and property value-at-risk within a 6-minute radius of the facility. The POS mentions that the facilities’ service area is susceptible to flooding (pg. 37) and indeed this is reflected in the fact that 4 of CommCare’s facilities rank above the 94th percentile for inland flood value-at-risk. Greenbriar in St. Tammany Parish has the highest risk: >97th percentile for value-at-risk and a Total risQ Score of 3.8, the probability of a hurricane making landfall here is ~1.5% each year. 

It’s not only physical risk that the CCRC are exposed to, many of them have Carbon Transition Risk. Baton Rouge, The Columns, Old Jefferson, Tioga, and Ridgecrest all have per capita CO2 emissions that rank above the 90th percentile in one or more economic sectors. In a future where carbon emissions are taxed, this means these areas are holding a relatively high liability and would expect to see higher energy costs, which can be especially impactful to seniors that live on fixed incomes. 

The Columns facility has the highest Social Impact Score (90) and three other facilities have Social Impact Scores above 80. The population near The Columns has a large portion of uninsured adults, just 78% have insurance which ranks near the absolute bottom in the nation. While CCRC are geared toward the retirement age community who have access to Medicare health insurance, private insurance leading up to retirement supports healthy aging. Worth noting is that among the CCRC group, 6 facilities have Persistent Health Obstacles Scores above 80. The Health Obstacles Score considers health insurance coverage and prevalence of an array of underlying conditions. Generally speaking, the score is a good measure of health outcomes for a community, and we see here that many of the areas served by these facilities have pervasive health issues.We acknowledge that CommCare is forthcoming with its disclosure on physical hazard risks (albeit lacking on discussion of carbon transition risk). Unfortunately discussing the problem isn’t enough to remedy the communities impacted by these nature disaster events. As Ida (and many other such events) has shown, the recovery period after these events can be long and painful, with many people still living without essentials such as electricity and a roof over their head more than one month after the storm.

Leave a Reply