Assessing Risks of School Reopening in Haverhill

Reopening schools in a world with COVID-19 presents both risks and uncertainties. We can estimate the probabilities of risks, such as COVID spread in the community, based on observed rates reported by our public health  officials. However, there are uncertainties, such as whether and how COVID-19 may spread in schools with the risk reducing methods of Haverhill’s hybrid model for reopening, for which we have little experience and probabilities are difficult to estimate. I have attached to this post slides with information on the risks of reopening Haverhill schools noting the remaining uncertainties that we can only learn with experience.

Assessing COVID Risks for Haverhill Schools 2020-08-27 COVID-19 Risks for Haverhill Schools 02

High schools are different, and so should be their reopening plans

We nearly all benefited from the varied learning experiences that our high schools provided – with a diverse set of classes with teachers, expert in their subject areas. However, the way such experiences are provided needs to be reconsidered in the era of COVID-19.

Recent evidence suggests that COVID-19 risks in high schools are greater, in part because the students are older, and in part because in rotating among classrooms students share space with many more persons in the course of a day. These risk could be reduced by changing the structure of student class assignments and/or schedules. While such changes would be difficult, they may be necessary to avoid uncontrolled outbreaks that could lead to disruptive school closings and/or extended periods of remote learning.

Haverhill Public Schools Opening Plan of August 11, 2020 calls for Haverhill High School students to attend classes in seven periods each in-person school day. The plan seems to presume that student class schedules and assignment to classrooms will be largely unchanged from prior practice. This presents an area of vulnerability to spread of the coronavirus.

Here are is the argument in a nutshell:

  1. We can expect spread of COVID-19 from the community to the schools.
  2. We need to contain these events in the classroom to avoid schoolwide and districtwide spread.
  3. The high school presents much greater risks of schoolwide spread.
  4. These risks could be greatly reduced by altering class rotation in scheduling.
We can expect spread of COVID-19 from the community to the schools.

In Haverhill community background rates for COVID though lower than many other states are currently higher than in many countries that have successfully opened schools. I am looking at the average of new cases per day in Essex County, as Haverhill statistics are based on fewer cases and produce more variable numbers. Also, many Haverhill residents engage in work, shopping or other activities beyond Haverhill’s borders that present COVID-19 risks.

How many new COVID-19 cases might we expect in a semester at HHS given our current local environment? We can only estimate this in very rough terms because we do not have good evidence-based estimates of COVID-19 spread  among children. New COVID-19 infection rates in Essex County are currently about 0.079 per 1,000 population per day in Essex County; Haverhill’s most recent (August 12) average daily incidence is .024 per 1000. We have about 2,400 students and teachers in Haverhill High School; assume 75% (1800) enroll in the hybrid in-person program. If student and teacher infections are similar to Essex County rates we might expect 0.079 x 1.8 x 30 days = 4.26 cases per month. [This would be 0.024 x 1.8 x 30 days = 1.30 with Haverhill numbers.] We might estimate about 61% of this or (0.8 to 2.6 cases per month) if most students only contract from a parent and do not work or have other activities outside the home and school. So roughly we might see on average about 1 to 4 COVID-19 cases per month arrive from the outside in each Model 2 cohort of the high school. If the rate of new cases in the community rises — perhaps due to local reopening, seasonal factors, or spread from states with higher infection rates – the rate could become higher.

We need to contain these events in the classroom to avoid schoolwide and districtwide spread.

Dr. Maddox’s told the School Committee that he was not much worried about individual “contained events.” Greater risks to health and educational disruption can come when the virus is spread more widely throughout schools and the district.

An infected student arriving in an elementary school a will be in contact with just one or two teachers and a small number of students in the elementary school classroom (of maybe 10 or 12 in the hybrid model). Even if the child remains asymptomatic and attends for several days the number of people will be small and the risk largely confined to one class room of students. Once discovered by testing or symptoms in one or more students the 10-12 students and their teacher can be quarantined without major disruption to the school. This is very likely a “contained event” such as Dr. Maddox noted.

The high school presents much greater risks of schoolwide spread.

It is important that school reopening plan recognize differences by age. Bill Gates, for example, has been speaking in favor of school reopening for students up to age 14 or 15. In Haverhill we have seen greater numbers of COCID-19 cases among those of high school age (14 to 17) than among other school age children. See chart.

Recent evidence suggests that children over age 10 can spread COVID-19 in a manner similar to adults. Compared to those in elementary schools, a similarly infected high school student may not only be a more effective spreader of COVID-19 but could be in contact with perhaps six or seven times as many students. These say 80 to 100 students will not all be in one class and they will in turn each be in contact with others in other classes, perhaps over several contagious but asymptomatic days.  So the virus could spread throughout the school (or cohort in the hybrid Model) in a matter of days. Not everyone would be infected but a majority of students in the high school (or cohort) could have at least on class, restroom, or hallway contact with an infected or suspected infected person.

This type of event, as soon as it is recognized, would probably be beyond capacity to timely identify trace and track hundreds of students. So it likely would necessitate shut down of the school for at least those in the infected cohort.  Such an event could be repeated as a newly infected students enter the school.

Such events are not just theoretical possibilities. As noted at the most recent school committee meeting, just this month North Paulding High School in Georgia which had to close after only one week of school when 9 students tested positive for COVID-19. HPS needs to find a way to avoid such school closures even if that means altering course assignments and schedules to avoid in school rotation among classes. Many of the largest reported COVID outbreaks in schools so far have been in high schools and middle schools.  This includes an outbreak in New Zealand (known for doing a good job on COVID control) that infected 96 persons, an outbreak in Israel where 153 students and 23 staff were infected, and a high school in France where 38% of pupils, 43% of teachers, and 59% of nonteaching staff had been infected. (Science Magazine, July 7, 2020).

These risks could be greatly reduced by altering class scheduling.

To avoid school-wide spread we need to keep kids in smaller groups (e.g. pods) that don’t interact with others. The challenge is high school students have multiple classes with different teachers. Here are a couple illustrative suggestions; you may come up with others. These methods, could be structured within the Hybrid Learning Model 2, would keep high school students involved in in-person learning in a single classroom with the same students all day, with teachers rotating among classes.

Option A: Group students in pods, each with the same core schedule. Students would choose among core course packages (e.g., Math, English Science History) generally grouped by level with some variation.  Those in academies could take core courses together. Core courses could be taught in-person, other classes such as languages and electives would be conducted on remote learning days scheduled for each cohort.

Option B: Intense Single-Subject Days. An alternative way to reduce contacts in high schools would be for students to stay in one class all day.  Students would have the same classes as traditionally but have only one subject and be in only one classroom each day in school. In the hybrid model students could meet in person for a full two days on one subject then meet the following week on another subject, rotating through classes week by week.

Option C: Sequential Courses. In a third such option, students might take classes sequentially one subject at a time. Courses would be taught much as an intensive summer school course over a few weeks.  Students would have only one subject for a month, rotating classes through all classes over the school year.

Any of these course scheduling approaches would reduce the number of high school students to come in contact with an infected student or teacher who enters the school building. Risk of spread, though perhaps higher among the older students, would be contained in a single classroom, as in an elementary school.


High schools are different and we need to address risk differently there. In implementing the Model 2 Hybrid plan for reopening Haverhill Public Schools should consider adapting student class schedules to reduce risk of school wide spread of COVID-19 in Haverhill High School. The choice may be between making such changes and disruptive closings throughout the school year.

It is possible, of course, that when students return to school community COVID-19 rates will be low enough, in-person attendance will be restricted enough, distancing will work well enough, masks will be effective enough, ventilation improvements will be complete enough, and compliance will be great enough that class schedule changes will not be needed. But all that seems uncertain at this point. In any case, the District should fully consider these issues and implement strong strategies to prevent schoolwide and districtwide spread that recognize different risks among the schools.

Reopening Schools and American Democracy

Real and terrible choices must be made to reopen schools. Very consequential choices with risks and uncertainty. The kinds of choices made in war and disasters. Choices to be made by those unprepared to make them.

Mayors, school boards, and committees elected to make local policies and budgets, are now asked to make very real and tragic choices about who is exposed to contagion and who is not, who will be educated and who will not.

Blame, if you will, an interconnected world, in which a Wuhan “wet market” can send disease around the world. Or blame underfunded international health organizations. Or blame an electorate who chose a national leader for his bluster, over competence and character. Or blame a President who crippled the federal response with weak appointments, distraction, and inaction. Or blame his enablers in Congress who discount science and government as solutions.

Blame or not, this year we ask our local leaders to make real and frightening choices. With tightened budgets and persistent coronavirus threats, students, families, and teachers will all face risks – many that could have been avoided.

If it is true that “Every nation gets the government it deserves,” then America must show it deserves better this November.

This post appeared as a letter to the editor of WHAV on July 24, 2020.

Schools reopening? Not so fast.

Around the world and around the country “Post-COVID-19” re-openings are in the news. What does this mean for reopening Haverhill Public Schools and other nearby districts where plans and budgets are now being developed for the summer and for the 2020-21 school year?

First, it is important to be clear that nowhere in the world is yet in a post-COVID-19 period. We will not be post-COVID-19 until an effective vaccine is widely available and this is most likely many months or even a year or more away. The continuing threat of COVID-19 needs to be effectively addressed in any school reopening plan.

Second, we need to realize that here in Haverhill, in Essex County, and in Massachusetts overall, we have been hit harder by COVID-19 than most other places around the country and around the world. Our own Essex County ranks 24th among the more than 3,000 counties in the United States in total number of confirmed COVID-19 cases. Around the world, many of the countries that are reopening schools have had many fewer COVID-19 cases and deaths on a per-capita basis (see Chart 1).  Compared to other communities, confronting the higher case levels in our area will take more resources for tracking and tracing, and it will probably take longer to control COVID-19 sufficiently to permit safe return to schools.

Chart 1

Third, high rates of new cases persist in our area. While daily new cases in Essex County have recently been declining, they are continuing at a higher per-capita level than in other Massachusetts counties and the state average (see Chart 2). As of May 13, Essex County had 1,359 new confirmed COVID-19 cases identified in the preceding seven days. At this rate, we would see more than 70,000 new cases per year in Essex County alone.  As businesses begin to reopen, the decline in new cases may not continue or new case counts may rise.

Chart 2

These statistics should be unsettling for cities and school systems, which have not received specific guidance on when to reopen schools. While we are nowhere near herd immunity, a seasonal or episodic decline in coronavirus infections and increased control of the virus might enable a return to school for the fall semester. On the other hand, if rates of new cases remain near current levels or resurge with business reopening, it might be advisable to continue remote learning for most or all students into the 2020-21 school year. Reopening safely will require that our local COVID-19 prevalence falls low enough so that (1) COVID-19 cases entering schools are rare and (2) the capacity to test, track, trace and quarantine grows high enough so that any outbreak is assured to be contained without spreading widely. 

A fourth factor to consider is whether detailed plans are in place and sufficient resources are allocated for schools to operate safely under new COVID-19 requirements. Local school districts must consider whether and how much to move away from the current remote learning model. Districts will need to implement procedures to minimize student contacts during transportation and throughout the school day. This could include testing of all students plus daily health screening, perhaps with temperature checks. School committees will need to make decisions and commit sufficient funds early enough so that facility and operational changes can be developed and put in place before students return to school.

Coordination is needed between schools and the public health system, particularly to facilitate community tracking and to ensure that the virus is not spread from the community to the schools or from the schools to the community. When new cases emerge it will be necessary to follow up each case, identify recent contacts, trace those contacts, notify them of COVID-19 exposure, and ensure 14-day quarantine. The new Massachusetts COVID-19 Community Tracking Collaborative (CTC) announced by Governor Baker on April 3 includes Partners in Health, which is hiring and overseeing a workforce for contact tracing within Massachusetts. It will be important to verify that their efforts are sufficient to detect new cases and to quickly contain any new outbreak in our area.

These difficult times will continue for some months ahead – until an effective COVID-19 vaccine is found and made widely available. There is no guarantee that this will happen any time soon. Our local school districts face difficult choices and implementation challenges for teachers, administrators, and staff. Shared sacrifices by workers, parents, and taxpayers will be required.  We will need to continue social distancing and refrain from high-contact shopping, entertainment, and businesses interactions. Parents will need to play a greater role in their children’s education. Taxpayers will need to support additional COVID-19-related budget requirements to keep our schools and our community safe.

Every generation faces challenges, and our current challenges can be overcome by maintaining our commitment to the education of our next generation as we contend with the health threats and economic adversities brought on by COVID-19.

Note: This piece also appeared in the Eagle-Tribune.

COVID-19 in Haverhill and Other Gateway Cities

(Updated to May 13th)

  • Gateway Cities have been hit harder by the COVID-19 virus than has Massachusetts statewide.
  • The number COVID-19 cases has been growing faster in Haverhill than in other Gateway Cities.
  • Haverhill is now in the middle of Gateway Cities ranking 12th among 26 Gateway Cities in cumulative COVID-19 cases per capita.

The Massachusetts Department of Public Health recently released COVID-19 case counts by city. Cumulative COVID-19 case per capita are on average 46 percent higher in the Massachusetts Gateway Cities than state wide (see Chart 1).

Chart 1

Among the 26 Gateway Cities Haverhill ranked 12th, up from 15th on April 29th and 19th on April 14th. This puts Haverhill just above the middle of this group of Massachusetts cities, many with diverse populations and facing economic challenges. While many factors affect case counts (including the amount of testing), moving up in the cumulative case-count-per-capita ranking raises some concern.

Chart 2 below shows the cumulative number of COVID-19 cases per 100,000 population, as identified by May 13, 2020.

Chart 2

Essex County COVID-19 Cases are Above State Average

(Updated to May 26, 2020)

Haverhill parents, teachers, school administrators and city leaders all need to make plans about education from home and reopening of schools closed due to coronavirus (COVID-19). Some have suggested that decisions to reopen businesses and schools might eventually be made county by county if patterns of coronavirus cases differ greatly among counties. This Benchmark Blog post looks at data available to date on coronavirus cases by county for the largest eastern Massachusetts counties.

This post considers cumulative and daily data on number of confirmed COVID-19 cases for the period for March 20 through May 26, 2020.  Case counts come from Massachusetts Department of Public Health (DPH). County population figures come from the UMass Donahue Institute. I computed cases per 1000 of county population (2018) for the largest Massachusetts counties and statewide. Keep in mind that counts reflect not only the incidence of the disease but also the amount of testing that is done. For May 26 the state DPH reported  only 4,920 tests. Cumulatively 17.1 percent of all 545,481 tests were positive.

Chart 1 shows Essex and the other non-Boston areas are on lower paths than Suffolk County. Essex County confirmed case counts per capita are now consistently well above the statewide average after tracking very near the state average up to April 2nd.  Essex County cumulative cases per capita are now 27.4 percent above the state average, and 35.5 percent above neighboring Middlesex County.

Chart 1

Chart 2 shows results for new cases reported daily. The greater day-to-day variation reflects the smaller numbers and possible day-of-the-week differences (fewer new cases tend to be reported on weekends) for testing and/or reporting. Suffolk County is tracking higher than other counties. As noted below, cases counts from April 13 to April 22 were under-reported to DPH by one laboratory company. The delayed cases were included in counts April 23 and 24, producing the spike in counts for those days.

Chart 2

Chart 3 presents these data for new cases as a seven-day moving average. This smooths out the day-to-day variation and avoids any day-of-the-week reporting effects. Part of the flattening of the curves April 13 to 22 is due to under-reporting in that period (see note below). The corrections on April 23 and 24th inflated case counts for those days and also inflates the 7-day moving average through April 30. So the apparent decline on April 30 and May 1 is largely due to reporting (the correction moving out of the moving average).

Monitoring updates to this chart may help identify when the COVID spread begins to slow in Massachusetts counties. A slowing of the spread of the virus would be indicated by a downturn in the 7-day moving average of new cases (not counting the drop due to reporting adjustment at the end of April).

Chart 3

Recently, in the first half of May, average new cases per capita declined particularly in Suffolk and Essex Counties, which had the highest rates in April. New cases are also trending downward statewide. Worcester County has seen the least decline and its new cases per capita are now tracking above the state average. With an apparent gradual decline in May, the rate of new COVID-19 cases in the larger Eastern Massachusetts counties are now closer together in the range of 10 to 20 daily new cases per 100,000 population.

There were 5,768 new confirmed COVID-19 cases in Massachusetts in the past seven days to May 26. This is down from 8,593 cases in the seven days preceding that period. These case counts, which reflect intensity of testing as well as prevalence of COVID-19, show the continued need for diligent social distancing and caution about reopening nonessential  businesses.  They do not yet point to an early end to COVID-19 threats in Essex and other eastern Massachusetts counties.

Note: The Massachusetts Department of Public Health (DPH) on April 24th noted a reporting error that reduced case counts for data from April 13 to April 23 due to a backlog in reporting cases from Quest Diagnostics. DPH provided revised case counts for that period for Massachusetts, but not for individual counties. The results reported here thus reflect under-reporting of COVID case counts for April 13 to 22 and increased case counts for April 24 to account for the previous under-counts