Nearly Half of Connecticut’s Audited Critical Incidents Went Unreported; State Considers Early-2027 Alerts

An alert promised later than a separate certification repair

Forty-eight of 101 critical incidents examined in Connecticut’s individualized home-support audit were not reported by providers. What’s the Scoop With Broach calculated that share—47.5%—and compared the full September 2026 audit with the state’s written corrective response. The proposed automated alert dashboard addressing missing incident reports has a different timeline from a separate provider-certification overhaul.

In its September 1 response, Connecticut’s Department of Developmental Services said it was reviewing the feasibility of an early-2027 launch for an emergency-room alert dashboard. It separately described a certification platform expected near the end of 2026, followed shortly by written procedures. Those are two systems with different functions and different degrees of commitment. The response does not establish that either has been fully implemented today.

The accountability question is whether a modernization announcement fixes the particular safeguard that failed. A platform that generates certification letters cannot, by itself, reconcile serious injuries against missing incident reports. Conversely, an emergency-room dashboard does not demonstrate that providers completed staff background checks or mandatory training.

Seventy incidents are not 77 incidents or 70 people

The HHS inspector-general audit used fiscal-year 2024 Medicaid data and worked with DDS to identify 101 hospital-related incidents that met the critical-incident standard. Seventy, involving 61 participants, were not properly reported or recorded. The remaining 31 in this reviewed group were not identified as deficient under that test.

The report lists four categories: 48 unreported by providers, 11 reported late, seven reported at an incorrect severity and 11 not recorded in the DDS system. Adding the categories gives 77, but a footnote explains the overlap: the seven incorrect-severity incidents are also included among the 11 not recorded. Subtracting those seven duplicates returns the 70 unique deficient incidents.

Thus the 48 never-reported incidents represent 68.6% of the 70 deficient incidents and 47.5% of all 101 reviewed incidents. These are descriptive ratios within the audit’s identified set, not an estimate that nearly half of every incident statewide went unreported. Nor does a reporting failure prove that each incident was abuse or that no care was delivered. DDS, at auditors’ request, determined whether the hospital visit represented a critical incident.

The reporting system was not fully implemented during the audited year. DDS said some incidents occur while support staff are absent and staff may not learn about them. Manual entry also allowed reported incidents to go unrecorded. Those explanations do not erase the required reporting, but they show why the audit concerns a chain of provider notification, severity coding and state recording rather than one interchangeable error.

Different denominators reveal other control problems and better results

The audit’s certification review covered 81 providers. Thirty-two lacked timely certification status letters: 17 eventually received letters late, 14 still had overdue letters as of July 2025, and one had no certification letter available. The state said providers had held annual meetings and completed checklists, but the checklists were not always sent to the unit that issued letters. Auditors could not verify all required certification work during the gaps. This is not evidence that every affected provider furnished unsafe care.

Annual participant-plan reviews performed much better in the selected sample: 118 of 124 were timely, or 95.2%. Five were late and one lacked evidence of review. The state said case notes supported a timely update in the documentation-missing case and reported 10,577 of 10,654 plans revised annually for a separate April 2024–March 2025 measure. That broader, differently timed measure should not replace the audit sample or be described as independent confirmation of all sampled records.

At five providers selected for onsite review, 106 of 132 employees lacked at least one timely required training. One provider accounted for 66 of those 106 employees—62.3% of the training findings—while employing 67 of the 132 reviewed staff. These selected providers were not a random statewide sample. The concentration shows where this particular set of findings accumulated; it cannot identify a statewide provider failure rate. Background-check and abuse-registry-check findings were separate and may involve the same employees.

The response and what remains unproved

Connecticut agreed with all four recommendations. Its response says the Pulselight claims-monitoring system was not fully operational during fiscal 2024; after launch, DDS explored extending its review to emergency-room encounters. The contemplated dashboard would combine visit information and alerts to flag claims without a current incident report. Provider orientation, record-retention terms and compliance monitoring are separate parts of its corrective approach.

The central findings concern oversight and increased risk, not a measured number of injuries caused by the deficiencies. We read the full audit, methods, overlap footnotes and state response, and checked the current official publication and work-plan records on October 15. The reviewed records do not establish a completed dashboard launch or independent testing of the corrections. This is original public-document analysis by What’s the Scoop With Broach, not firsthand observation of supported living services. It identifies the distinct repair timelines and reconciles the incident counts so that promised changes can be judged against the safeguard each is meant to restore.