Results

Outbreak Characteristics

During 2013–2017, a total of 7 HHC-associated iGAS outbreaks were identified in England; during January 1, 2018–August 23, 2019, a total of 10 HHC-associated iGAS outbreaks were identified (Figure 1). In these 10 outbreaks, 96 iGAS cases and 28 attributable deaths (case-fatality rate 29%) were reported. Outbreaks ranged from 2 to 39 (median 7) iGAS cases; case-level data and results of HHCW screening for 1 outbreak (outbreak number 10) were unavailable (Table 2, Table 3).

Enlarge

Figure 1.  Annual number of home healthcare–associated invasive group A Streptococcus (iGAS) infection outbreaks reported to Public Health England, January 1, 2013–August 31, 2019. A total of 17 outbreaks occurred during this timeframe, but outbreaks sharply increased during 2018–2019.

The median age of case-patients was 83 (range 42–100) years; 68% of cases were among female patients and 32% among male patients. Among 96 cases, 92 (96%) patients received nursing care administered by HHC services. Of the 4 cases that did not receive direct HHC care, 2 were household contacts of patients receiving HHC and neither had an identified GAS infection at the time. An epidemiologic link to HHC was not established for the other 2 cases, but those 2 were linked to other outbreaks by WGS.

Among 5 outbreaks with recorded wound swab sample results, GAS was cultured from 104 case-patients (range 1–95 cases per outbreak). The number of bacterial swab samples taken in these outbreaks was not documented by investigating teams, and available data did not enable distinction between GAS carriage and noninvasive infection (Table 2).

Outbreak Identification

Nine outbreaks were identified through statutory notifications of individual iGAS cases to local HPTs; 1 outbreak (outbreak 4) was identified through WGS at the RVPBRU Streptococcal Reference Laboratory. The median time between first identified case and the date the outbreak was declared was 40 days (range 3–517 days), but these data were not available for outbreak 10. Some cases were identified retrospectively when investigation teams reviewed previously notified iGAS cases of the same emm type to reinvestigate a link to HHC (Figure 2).

Enlarge

Figure 2. Timeline of cases in 9 home healthcare–associated invasive group A Streptococcus (iGAS) infection outbreaks, England, January 1, 2018–August 31, 2019. Vertical black line indicates date that outbreak was declared. Diamonds indicate day of initial detection of iGAS cases: blue diamonds represent patients that survived, red diamonds patients that died. Data from outbreak 10 (39 cases, 15 deaths) were not available.

Six outbreaks were caused by S. pyogenes type emm1 or emm89, the 2 most common iGAS-causing emm types circulating in England during this period. Among the remaining 4 outbreaks, 2 were caused by emm94, 1 by emm87, and 1 by emm44. WGS was performed for 6 outbreaks involving emm1 (n = 2), emm89 (n = 3), and emm94 (n = 1) to establish whether cases of common emm types with epidemiologic links constituted an outbreak. Outbreak 10 (emm44) was sequenced because of the substantial number of cases and long duration (Table 2).

In the 6 outbreaks of common emm types (emm1, emm89, emm94), WGS confirmed that epidemiologically linked cases formed a genomic cluster in each outbreak. In 3 of these outbreaks, WGS identified ≥1 case of the same emm type with epidemiologic links to the outbreak that did not cluster with the other cases, enabling exclusion of the case from the outbreak. In 2 outbreaks, WGS confirmed that 2 sequential cases diagnosed >5 months apart but cared for by the same HHC team formed a genomic cluster and were likely part of the same outbreak. None of the sequenced outbreaks had close genomic relationships with each other, indicating each was a distinct outbreak.

One outbreak (outbreak 4) was not initially recognized by the local HPT but was identified by the reference laboratory from a set of local WGS controls used to investigate another HHC-associated iGAS outbreak (outbreak 9) (Table 2). The discovery of outbreak 4 revealed a separate emm89 iGAS in patients cared for by a single HHC team. Outbreak 4 involved 7 cases and 2 deaths over a period of 388 days, and the last case was notified 74 days before the outbreak was identified; no further cases were identified in the 60 days after the outbreak was identified. Although case-patients were cared for by a single HHC team, the epidemiologic link between cases was not identified earlier because the outbreak involved emm89, a common type; long intervals passed between sequential cases; and the HPT did not routinely ask about HHC exposures.

Outbreak Duration

Duration of outbreaks varied greatly. The median time between specimen collection from the first and last identified case in each outbreak was 199 days (range 3–507 days). Long intervals often passed between cases (median 20.5 days, range 1–225 days) (Figures 2, 3).

Enlarge

Figure 3. Intervals between sequential invasive group A Streptococcus (iGAS) cases in 9 home healthcare–associated outbreaks, England, January 1, 2018–August 31, 2019. Data from outbreak 10 were not available.

In outbreaks 2, 4, 8, and 9, the last recognized case occurred before the outbreak was formally declared, and these outbreaks might have self-terminated after HHC teams instigated improved infection control and before the HPT became involved (Figure 2). Specifically, outbreaks 4 and 9 occurred in a region with a large concurrent HHC-associated iGAS outbreak in which HHC services had recently reviewed their infection control procedures. In the other 6 outbreaks, a median of 130 days (range 31–181 days) passed between outbreak declaration and the last identified case.

Once outbreaks were identified, time to link outbreaks to HHC was often delayed. Among 48 case-patients for whom place of residence was documented, 17 (35%) lived in residential care but also received HHC services. Transmission within the residential care facility initially was investigated before further cases were identified outside this environment and HHC links were explored.

Outbreak Investigation

Investigating teams performed network analyses during outbreak investigations through records provided by HHC teams. These investigations did not identify a single HHCW in contact with all case-patients during the 7 days before symptom onset. HHCWs visited up to 20 patients per day, and multiple HHCWs might visit a patient each week, making investigation complex. In 5 outbreaks, ≥1 HHCW described symptoms suggestive of GAS before or during the associated iGAS outbreak. In addition, 8/10 OCTs reported difficulty obtaining information from HHC teams because of poor record keeping and time pressures on already overstretched services.

After network analyses, HHCWs were screened with throat swab samples for bacterial culture in all 10 outbreaks. The aim of screening was to identify HHCWs who might have acted as a common source and posed an ongoing risk to patients. In the 9 outbreaks for which data were available, a total of 411 HHCWs were identified for screening and 366 were screened by throat swab. A median of 22 (range 3–160) HHCWs were screened per outbreak. A single (0.36%) throat swab sample cultured GAS but unfortunately was not typed. In 7 outbreaks, any reported wounds or skin breaks among HHCWs were screened for GAS by swab and culture, but all were negative. In 3 outbreaks, a few HHCWs with negative throat swab samples but strong epidemiologic links to cases were screened with swab samples from piercing sites, perineum, and vagina; none were positive. The logistics of screening HHCWs in the community were complex, predominantly because of inadequate occupational health provision (6/8 outbreaks) and delays of up to 6 weeks between the decision to screen and commencement of screening. In addition, HHCW screening involved associated sensitivities, including concern about the use of screening to attribute blame and potential personal shame if swab samples were positive.

In 3 outbreaks, patient wounds were systematically screened for GAS carriage. In the 2 outbreaks with data available, 107 patients were screened but no GAS-positive samples identified. Although full data are not available for the third outbreak, GAS carriage and infection was detected in a small proportion of patients. In 7 outbreaks, patient wound screening was not systematically performed, but in 4 of these outbreaks HHCWs were encouraged to send swab samples from any wound with suspected infection. Although the number of swab samples sent for this indication is unknown, 6 swab samples from 2 outbreaks tested GAS-positive, but these were not emm typed, so they cannot be directly linked to other outbreaks.

In 2 outbreaks, environmental screening was performed. Bacterial swab samples were taken for culture from communal and storage areas at the HHCW base and from items that were difficult to clean, including portable electronic devices (e.g., tablets or smart phones), equipment, bags, blood pressure cuffs, and Doppler machines. Although the total number of swab samples taken was not recorded, a single swab sample taken from the handle of an equipment bag cultured GAS-positive, and subsequent WGS confirmed it to be the outbreak strain.

Source and Transmission Mode

The sources and modes of transmission were not definitively established in any outbreak. The common hypothesis among investigating teams was that GAS was transmitted between colonized or infected patients and HHCWs and that numerous possible transmission events caused each outbreak. The role of fomites was unclear, but teams recognized the challenges associated with adequately decontaminating HHCW equipment in the home environment.

Infection Control Methods

Infection control procedures were reviewed in each outbreak. Recommendations included infection control training for HHCWs and enhanced cleaning of HHCW bases and equipment storage areas in their cars. In 5 outbreaks, investigators noted that HHCWs carried equipment that was difficult to clean, such as fabric bags, portable electronic devices, and Doppler machines. This finding led to replacing fabric bags with impermeable, surface-wipeable bags (n = 3) or plastic, wipeable crates (n = 1), along with developing standard operating procedures for cleaning equipment that was difficult to decontaminate (n = 2). After outbreak 10 was identified, HHCWs were given disposable long aprons to wear during wound care procedures.

In 7 outbreaks, HHCWs were treated with antimicrobial drugs, which were intended to decolonize staff with potential occult carriage and interrupt transmission. In 6 outbreaks, HHCWs who had direct contact with a case-patient were initially treated with a 10-day course of penicillin V (median 2 [range 1–3] HHCWs per outbreak). When further cases occurred in 5 outbreaks, mass penicillin V prophylaxis for HHCWs was advised by the OCT and administered. In 4 outbreaks for which data were available, 139 HHCWs received prophylaxis (median 26 [range 22–65] per outbreak). In 3 of these outbreaks, no iGAS cases were notified after mass prophylaxis. HHCWs voiced opposition to antimicrobial drug prophylaxis in 3 outbreaks because of perceived lack of need after negative screening and concerns about antimicrobial resistance. In outbreak 1, the HPT directly engaged with HHCWs through presentations and discussions to achieve reasonable coverage and compliance with antimicrobial prophylaxis. Overall, HHCW compliance to antimicrobial prophylaxis is unknown.

Patients whose wounds cultured GAS-positive were treated with antimicrobial drug therapy. Mass antimicrobial prophylaxis was not administered to patients in any outbreak.