Health Information Technologists and Medical Registrars

• Paper charts1912–1966

  Paper charts, card index systems, and disease indices

  For the first four decades of the profession, every medical record was a paper document: handwritten physician notes, typed discharge summaries, nursing observations, lab slips. The medical record librarian's primary tools were card files and ledger books: a disease index (listing all patients by diagnosis), an operation index (by procedure type), and a physician index (by attending clinician). Records were filed in folders in a terminal-digit filing system to distribute retrieval traffic evenly across the stacks. Retrieving a chart for a readmission or a researcher meant a physical walk to the shelves. The ICD system, which the United States adopted in codified form beginning with the 1949 ICD-6 revision, provided the first standardised vocabulary for the disease index, though not yet for billing.

• Medicare1966–1983

  Medicare and Medicaid documentation requirements (microfilm, early mainframes)

  The Social Security Amendments of 1965, which created Medicare and Medicaid, transformed the medical record from a clinical document into a financial instrument. For the first time, a hospital's reimbursement depended directly on the accuracy and completeness of the records its medical record department produced. The profession grew rapidly in the late 1960s and 1970s in response. Microfilm became the storage medium of choice for completed records, reducing physical space requirements. Large teaching hospitals began experimenting with mainframe-based master patient index systems in the 1970s, but most medical record departments remained paper-and-microfilm shops. The 1979 federal mandate to use ICD-9-CM (the Clinical Modification developed specifically for US billing) for all Medicare and Medicaid claims was the first requirement that every hospital employ credentialed coders capable of translating physician documentation into standardised diagnostic and procedure codes.

  Effect on the work

  The Medicare and Medicaid reimbursement link created a direct financial incentive for hospitals to invest in credentialed medical record professionals. AHIMA membership grew substantially through the 1970s; the Accredited Record Technician (ART) credential, created in 1953, became a hiring requirement at many hospitals.

• Diagnosis-Related Groups1983–2000

  Diagnosis-Related Groups and encoder software (Medicare Prospective Payment System)

  When Congress amended the Social Security Act in April 1983 to replace cost-based Medicare hospital reimbursement with a fixed-rate Diagnosis-Related Group (DRG) system, the financial stakes of accurate coding became acute. A hospital was now paid a single predetermined rate for each Medicare admission based on the DRG assigned by the coder, regardless of actual costs. An incorrect principal diagnosis or a missed secondary condition could mean tens of thousands of dollars in underpayment or, if codes were inflated, potential fraud liability. Health information departments acquired dedicated encoder software, first running on DOS and later on early Windows systems, that prompted coders through the ICD-9-CM decision trees and automatically assigned DRGs. Studies of early DRG coding found error rates of roughly 20% in the first years of the program, which drove investment in coder training and quality auditing as a core HIM function.

  Effect on the work

  The DRG era was the single biggest driver of HIM employment growth in the profession's history. Hospitals hired coders aggressively from 1983 onward; the ART credential became near-mandatory for hospital coding positions. The profession's financial leverage in the hospital grew substantially, as a proficient coder could legitimately recover millions in underpayment annually by capturing all documented secondary diagnoses and complications.

• Enterprise Electronic Health Record sys…2000–2015

  Enterprise Electronic Health Record systems (Epic, Cerner, HITECH mandate 2009)

  The Health Information Technology for Economic and Clinical Health (HITECH) Act, signed into law on February 17, 2009 as Title XIII of the American Recovery and Reinvestment Act, allocated approximately $25.9 billion to drive adoption of certified electronic health record systems by US hospitals and physicians. By 2015, 96% of hospitals and 87% of office-based physician practices were using EHR systems. For health information professionals, EHR adoption was both a professional windfall and a job-content revolution. The transition required enormous human effort: every hospital needed staff to configure EHR templates, map legacy paper codes to electronic equivalents, train clinicians, validate data migration, and implement the new coding workflows within the system. The ICD-10-CM transition, which the US completed on October 1, 2015, tripling the number of available diagnostic codes from roughly 13,000 (ICD-9-CM) to 68,000 (ICD-10-CM), required years of coder education and created a temporary wave of additional hiring to manage the changeover.

  Effect on the work

  BLS OEWS data for the predecessor code 29-2071 shows the workforce growing from roughly 167,000 in 2010 to approximately 200,000+ by the mid-2010s as EHR implementation and ICD-10 transition drove demand. The 2018 SOC revision that split the category into 29-9021 (technologists) and 29-2072 (specialists) makes longitudinal comparison difficult.

• ICD-10-CM / ICD-10-PCS2015–2022

  ICD-10-CM / ICD-10-PCS, computer-assisted coding (CAC), and cloud-based EHR analytics

  The October 2015 US transition to ICD-10-CM (diagnoses) and ICD-10-PCS (inpatient procedures) was the largest single structural change in clinical coding since the 1983 DRG system. The new code set was five times larger than ICD-9-CM and introduced greater specificity in laterality, encounter type, and clinical detail. Early computer-assisted coding (CAC) tools, using rule-based natural language processing to suggest ICD codes from physician note text, became a standard feature of major EHR vendor platforms. CAC did not replace coders: it served as a first-pass suggestion engine that a human coder reviewed, accepted, rejected, or modified. The practical effect was to shift coder time from routine lookups toward complex multi-condition cases and query management. Cloud-based analytics platforms from vendors including Health Catalyst, Optum, and 3M gave health information departments access to population-level data analysis tools, expanding the role toward outcomes reporting and quality metrics.

• Generative AI clinical documentation2022–present

  Generative AI clinical documentation and autonomous coding assistants (ambient scribes, LLM-assisted coding)

  Ambient AI scribes, combining automated speech recognition and generative large language models, have emerged as the most disruptive near-term technology for health information professionals. These tools listen to patient-physician conversations and draft structured clinical notes automatically, reducing the quality and quantity of physician documentation that coders work from. In parallel, AI coding assistants have moved beyond the rule-based CAC of the 2010s toward LLM-based systems that claim fully autonomous ICD and CPT code assignment. The honest picture as of 2026 is that LLM-based coding systems achieve below 50% exact-match accuracy in studies using complex inpatient cases, compared to trained human coders who achieve 80-98% accuracy depending on case complexity. The hybrid model, AI as a triage and first-pass engine with human review of low-confidence and high-complexity cases, is the mainstream enterprise approach. The net employment effect through 2034 remains positive: the BLS projects 15% growth for 29-9021, citing the continued expansion of digital health data that requires governance, quality assurance, and compliance oversight that AI systems cannot yet provide without human validation.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.