99232 vs 99233: Subsequent Inpatient Levels

No, and this is the single most widely repeated error about these codes. The 2023 CPT descriptor states that "when using total time on the date of the encounter for code selection, 35 minutes must be met or exceeded" — a floor, with no upper bound written into the code. The reason a range gets reported is that 50 minutes is where 99233 begins, so people infer a 35-49 band. The practical effect is usually the same, but the language matters when you are defending a note: the code says met or exceeded.

No. Location does not set the level. CPT is unusually direct here: for hospital inpatient and observation care, the problem addressed is the problem status on the date of the encounter, which may be significantly different from admission, and may not be the cause of admission or continued stay. A patient stabilised in the ICU whose management today is routine does not meet the high descriptors, and a ward patient who deteriorates today can. Score today's problem, not the bed.

No. The 2023 revision rewrote 99231-99233 and their guidelines; the interval-history requirement and the old two-of-three key-components method went with it. The codes now require "a medically appropriate history and/or examination" and are levelled by MDM or by total time. A great deal of guidance still circulating online — including forum answers that rank highly in search — describes the pre-2023 rules with detailed histories and exams. Check the date on anything that mentions key components.

Effective 1 January 2023, CPT deleted hospital observation services codes 99217-99220 and the subsequent observation care codes 99224-99226, and merged observation into the hospital inpatient family. Initial care is now reported with 99221-99223 and subsequent care with 99231-99233 regardless of whether the patient is in inpatient or observation status. CPT also provides that a stay including a transition from observation to inpatient is a single stay, so the transition itself does not restart the initial-service clock.

Two of three elements, as everywhere else. On problems, high needs a severe exacerbation, progression or side effect of a chronic illness, or a threat to life or bodily function — a complex but stable patient usually does not clear that. On data, high needs two of the three categories where moderate needs one, and in hospital work the second category is frequently available: an independent interpretation of imaging read by radiology, or a documented discussion of management with a consultant outside your group. On risk, the setting-specific high example is a decision to escalate hospital level of care. Complexity of the patient is not the test; complexity of today's decision making is.

Under CPT, prolonged code 99418 becomes reportable for subsequent care at 65 minutes — the 50-minute base for 99233 exceeded by a full 15 — in complete 15-minute units only. Medicare declined to adopt 99418, on the stated ground that its billing instructions would create administrative complexity and potentially duplicative payments, and created HCPCS G0316 instead. G0316 is billable with 99233 at 80 minutes. Beware a trap in CMS's own current Evaluation and Management MLN booklet: its prolonged-services table lists 65 minutes for 99233 under a column headed "time threshold to report prolonged," but the surrounding text requires at least 15 minutes beyond that figure. The billable number is 80, not 65, and a page that copies the table without the text is wrong by a full unit. Both codes require the visit to have been levelled on time alone, and G0316 may not be billed with discharge codes 99238 or 99239.

More than any other E/M code CMS reports. In the current CERT data, 99233 carries a 22.7% improper payment rate (95% CI 18.7-26.7%) and $502m in projected improper payments — the highest on both measures, and more than five times 99232's 4.1%. Incorrect coding accounts for 57.7% of it, so these are largely notes that exist and are complete but do not support high MDM. Subsequent hospital visits as a category run at 13.8% against 5.3% for established office visits. The historical direction of travel is the same: OIG found in 2012 that physicians increased their billing of higher-level E/M codes across all service types between 2001 and 2010, and on CMS utilisation data 99233's share of subsequent visits rose from 32.3% in 2019 to about 37% from 2021 onward while 99231 fell from 8.9% to 6.6%. None of that makes a well-documented 99233 wrong — but it does mean the code is reviewed against a sceptical prior, and that the note has to show today's severity rather than the admission's.

As one service. CPT treats a physician or QHP of the exact same specialty and subspecialty in the same group practice as the same clinician for these codes, and subsequent care is reported per day. If you are levelling on time, only distinct time may be summed — where two or more individuals jointly meet with or discuss the patient, only one individual's time is counted. Advanced practice nurses and physician assistants working with physicians are treated as the same specialty and subspecialty as the physician.