Research program · 130 proposed investigations · 12 clusters · compiled
The KPSGILL Research Program
What the written rule says, against how the decision is actually made
One hundred and thirty investigations this institute intends to run, each stated as a question a public record can answer. Every item is proposed work, not a finding.
Compiled by Kanwar Partap Singh Gill, MD · compiled
The premise
This institute has spent its first phase building horizontal coverage: a policy library, standing regulatory trackers, an enforcement ledger, a legislation desk, living dossiers, a relationship graph and an event timeline. That work continues. It is no longer where the greatest unmet need is.
The next frontier is empirical institutional analysis: take public records that already exist but are effectively impossible for an ordinary reader — or an ordinary physician — to synthesise, structure them into datasets, and ask the questions the issuing bodies do not answer about their own output.
Another hundred explanatory essays would describe the machinery more thoroughly. They would not tell a physician what the machinery actually does.
Also on this page: the project that starts first · the twelve clusters · what public records can and cannot support · sequence · the point of view · how an item becomes a dataset
Every item below is written as a question with a lifecycle state of PROPOSED. None is a finding. Proposed routes appear as plain text rather than links, because an address that does not yet serve a page has no business looking like a destination. Items are numbered as they were proposed, so the program can be discussed by number.
The project that starts first
California Medical Discipline Observatory — what physicians are accused of, what the evidence establishes, what consequence is imposed, and on what stated reasoning. Its first dataset is California alcohol-related physician discipline.
The research question
Why can an apparently similar off-duty alcohol-related event result in a public reprimand for one California physician and years of monitored probation for another?
The observatory does not open with the claim that discipline is arbitrary. That proposition is not established by the public record, and it is not the interesting question. The published disciplinary guidelines set out recommended ranges; actual dispositions in comparable-looking matters can sit well below those ranges, and published administrative decisions sometimes say why — elapsed time, absence of a demonstrated pattern, abstinence, treatment, no diagnosed disorder, and public protection rather than punishment as the stated object. That is not evidence of favouritism. It is evidence that a guideline range does not mechanically determine an outcome, and that what reaches the decision-maker may matter more than the label on the charge.
- Code the charge selection
- Code the facts the decision-maker found
- Code representation and procedural route
- Code the evidence that reached the decision-maker
- Code the reasoning given for any departure from the published range
- Code the final action and its burden
- No named individual outcome published from a secondary summary
- Every record verified against the issuing body’s published document
- Allegation distinguished from finding in every field
- No correlation reported as cause
If, after controlling for charge selection, findings, representation, procedural route and the evidence actually presented, unexplained divergence remains, this institute will say so and show the working. If instead the divergence is explained by developed rehabilitation evidence, expert testimony and contested adjudication, that is the more consequential finding: the nominal range tells only part of the story, and the practical outcome depends on what the decision-maker was given. The codebook is published: what counts as a case, which document governs each field, and what the numbers are permitted to mean. Related published analysis: the standard of proof, how the hearings work, the proposed decision, why a complaint is not a finding.
The program
130 proposed investigations in 12 clusters. Rendered from the program record, which is the system of record for this page.
Medical discipline observatory
What the Board actually does when comparable cases reach disposition — as distinct from what the machinery says it will do.
Medical Board decisions observatory PROPOSED
Structured records for every publicly available California final decision: identity, procedural route, conduct category, clinical nexus, aggravation, mitigation, outcome, probation burden.
Sanction variation project PROPOSED
Why can an apparently similar off-duty alcohol-related event produce a public reprimand for one California physician and years of monitored probation for another?
One conduct, different penalties PROPOSED
Within a single conduct category, which factual, procedural and evidentiary variables track the difference in outcome?
Penalty lab: guideline versus disposition PROPOSED
For each violation, the published guideline range, the actual disposition, the size of any departure, and the reason the decision-maker gave for it.
ALJ decision laboratory PROPOSED
Across the health boards at the Office of Administrative Hearings: which mitigating and aggravating factors do administrative law judges repeatedly credit?
Board versus ALJ observatory PROPOSED
How much deference does a board actually give its administrative judges — adopted unchanged, sanction increased, sanction reduced, findings changed, non-adopted?
Defence variables observatory PROPOSED
Do publicly observable factors — representation, contested hearing versus stipulation, expert and treating-clinician evidence, documented rehabilitation — track departures from the default range?
Charging discretion PROPOSED
Complaint, investigation, referral, accusation, amendment, settlement demand, hearing: how do charging choices change the penalty that becomes available?
The impairment allegation PROPOSED
How does an alcohol-related matter charged as a conviction alone differ in consequence from the same matter charged with an impairment or practice-under-influence count?
Mitigation that is actually credited PROPOSED
Every mitigating factor expressly credited in a final decision, extracted and counted rather than assumed.
Aggravation that changes the outcome PROPOSED
The mirror project: which aggravating findings recur in the decisions that end in revocation or actual suspension?
Probation is not one penalty PROPOSED
Nominal sanction versus real-world burden: obligations, testing, monitors, practice restrictions, disclosure, duration, cost.
Discipline cost observatory PROPOSED
Investigation and enforcement costs requested and awarded, by offence and by procedural path.
Rehabilitation versus punishment PROPOSED
When a regulator invokes public protection, what prospective risk has actually been demonstrated?
Proportionality across professions PROPOSED
For comparable off-duty conduct, how does physician discipline compare with lawyers, nurses, dentists, pharmacists and psychologists?
Fifty-state discipline atlas PROPOSED
One hypothetical set of facts, fifty licensing systems: burden of proof, tribunal, penalty guidelines, impairment rules, disclosure, judicial review.
Board transparency index PROPOSED
Not due process — visibility. What can the public actually see: accusations, complete decisions, proposed decisions, votes, recusals, historical records, bulk data?
Digital permanence of discipline PROPOSED
A sanction has three lives: the legal order, the licensing record, and the search result. When the statute says a record comes down, does the internet agree?
Decision search PROPOSED
A structured search over public decisions: conduct plus outcome plus reasoning, rather than a list of PDFs.
Precedent without precedent PROPOSED
Administrative decisions are largely non-precedential, yet penalty custom develops. Where does custom substitute for rule?
Discretion, decision rights and the administrative state
Where written law ends and institutional judgement begins.
Who actually decides PROPOSED
Decision-rights maps for formulary, privileges, prior authorisation, licensure, residency selection, certification, closure, device clearance, guidelines, payment and scope.
Discretion institute PROPOSED
Prosecutorial, regulatory, credentialing, medical-necessity, program-selection, certification and enforcement discretion, studied as one subject.
Health administrative state observatory PROPOSED
Notice and comment, guidance, enforcement discretion, adjudication, deference after the recent doctrinal shifts, appointments, emergency authority.
Licensing and the constitution PROPOSED
Property and liberty interests in a licence, procedural and substantive due process, compelled speech, vagueness, retroactivity, subpoena authority.
Employee and licensee at once PROPOSED
The employed physician answers to employer, board, medical staff, payer, federal programs, carrier and certifying body. Where do those masters conflict?
Settlement pressure PROPOSED
What rights are waived, and what record survives, when a contested hearing is traded for a stipulation?
Public-protection evidence project PROPOSED
A standing method: whenever public protection is invoked, ask what risk was demonstrated and how.
Risk versus punishment PROPOSED
Do sanctions correlate with any measured probability of recurrence or patient harm?
Institutional power atlas PROPOSED
For each institution: who appoints, who funds, who can overrule, who audits, who sues, and who hears the appeal.
Revolving door register PROPOSED
Documented transitions between regulators, agencies, systems and payers. Movement recorded; motive not imputed.
Professional standing and the career cascade
How one event propagates through institutions that never speak to each other.
Whistleblower intelligence PROPOSED
False Claims Act actions, state protections, retaliation, patient-safety reporting, peer-review retaliation, compliance channels.
Medical staff democracy PROPOSED
Do organised medical staffs retain independent governance — bylaws, elections, executive-committee authority, governing-body override, exclusive contracts?
Bylaw library PROPOSED
Publicly available medical-staff bylaws compared on hearing rights, summary suspension, automatic relinquishment, conduct policies and panel composition.
NPDB consequence observatory PROPOSED
What a report actually does downstream: credentialing question, payer question, licensing disclosure, employment consequence, underwriting.
Credentialing exclusion filters PROPOSED
The screens applied before anyone reads an application: years since training, certification, claims history, gaps, probation, visa, volume.
Career fragility map PROPOSED
One event traced through every institution that will eventually ask about it.
Regulatory cascade observatory PROPOSED
Criminal court, board, hospital, federal data bank, payer, employer, carrier, immigration authority, other states — from a single fact.
Reputation and due process PROPOSED
Publication of accusations, arrests, peer-review outcomes and discipline, and the distance between accusation and adjudication.
Convictions and the professional map PROPOSED
Arrest through prosecution, mandatory reporting, board investigation, institutional reporting, credentialing and immigration consequence — as system mechanics.
Rehabilitation institute PROPOSED
How health regulation treats past addiction, mental illness, disability, conviction and prior discipline. When has enough rehabilitation been shown?
Second-chance licensing PROPOSED
State-by-state treatment of the rehabilitated clinician.
Ageing and cognitive screening PROPOSED
Age-based screening, competency testing, institutional policy, disability law, and the evidence for the tests being used.
Professionalism observatory PROPOSED
Who defines professionalism, and what process attaches to standards written as adjectives?
Off-duty conduct and medicine PROPOSED
When does private life legitimately become professional regulation — across conviction, speech, domestic dispute, accusation, recorded conduct and financial conduct?
Viral governance PROPOSED
Institutional due process under reputational pressure: what happens to process when the audience arrives first.
Physician speech PROPOSED
Scientific disagreement, misinformation statutes, employer restrictions, testimony, advocacy, and the First Amendment questions underneath.
Portability after discipline PROPOSED
What happens to a physician disciplined in one state who applies in another.
Compact consequence PROPOSED
How discipline propagates through interstate licensure compacts.
Evidence, expertise and the record
Who gets to say what happened, and on what basis.
Malpractice outcome analysis PROPOSED
Specialty, allegation, injury, standard-of-care issue, disposition where public, and the damages framework — structured, not advertised.
Guidelines in litigation PROPOSED
Where clinical guidelines become evidence, argument, safe harbour or cross-examination material.
Medicine and criminal law PROPOSED
Prescribing prosecutions, fraud, controlled substances, death prosecutions, certification and documentation offences.
The complaint economy PROPOSED
Who may complain, anonymously or not, across boards, peer review, grievances, payers and hotlines — and what a complaint sets in motion.
Anonymous allegation and due process PROPOSED
When can an anonymous allegation carry serious professional consequence?
Expert accountability PROPOSED
Qualification, conflicts, specialty matching, compensation, review standards and accountability for erroneous opinion.
The consultant system PROPOSED
Reviewer availability, standards, conflicts, timeliness, disagreement and anonymity in agency clinical review.
Delay, cost and access to administrative justice
Process that is nominally available and practically out of reach.
Board delay observatory PROPOSED
Complaint to investigation to accusation to hearing to decision, by case type. Delay is itself a consequence.
Administrative justice delay index PROPOSED
The same measurement across licensing boards.
Cost recovery observatory PROPOSED
Whether cost recovery discourages the exercise of hearing rights.
Rulemaking, influence and who wrote the rule
Tracing authorship of rules that are presented as authorless.
Society influence observatory PROPOSED
Positions taken by professional, hospital, payer and consumer organisations, compared with the outcome.
Who wrote the rule PROPOSED
Statute, agency draft, comments received, final text — showing which recommendations survived.
Comment-letter influence PROPOSED
Recurring institutional positions across major federal and state rulemakings.
California power map PROPOSED
Who regulates whom: executive, legislature, consumer affairs, the health departments, insurance regulators, the Attorney General, systems, plans, societies, unions, counties and capital.
The hidden corporations PROPOSED
Who is actually paid when a patient sees a doctor.
Contracting laboratory PROPOSED
Publicly available payer, hospital and government contracts, read as instruments of control.
California places and the safety net
Where policy meets a county line.
County health governance PROPOSED
Fifty-eight counties: hospitals, behavioural health, public health, jails, emergency services, Medi-Cal delivery, indigent care.
Safety-net systems observatory PROPOSED
Health centres, county systems, rural clinics and public hospitals: governance, funding, workforce, access.
Health-centre governance PROPOSED
Federal grant conditions, board composition requirements, liability coverage, drug pricing, payment methodology and clinical authority.
Drug-discount programme observatory PROPOSED
Covered entities, contract pharmacies, manufacturer disputes, litigation, and the patient-benefit question.
Nonprofit hospital accountability PROPOSED
Exemption against community benefit, executive pay, charity care, pricing, collections and service closure.
Compensation and community benefit PROPOSED
Public filings read side by side.
Charity-care observatory PROPOSED
What hospitals promise against what they provide.
Facility fee atlas PROPOSED
The same billed service, mapped by site of care and ownership.
Debt collection atlas PROPOSED
Which systems sue patients, under which written policy.
Coverage, payers and middlemen
The layer patients are billed by and cannot see.
Appeals outcome database PROPOSED
Denial, appeal, external review, reversal — wherever public data permit the comparison.
Medical necessity criteria library PROPOSED
Published payer criteria for the same service, compared.
Payer policy change tracker PROPOSED
Clinical policy updates that function as regulation without rulemaking.
Prior authorisation burden PROPOSED
By drug and by procedure, concretely.
Formulary churn tracker PROPOSED
Which medicines change tier or exclusion, and when.
Benefit-manager ownership graph PROPOSED
Manufacturer, benefit manager, insurer, pharmacy and specialty pharmacy in one diagram.
Practice acquisition tracker PROPOSED
Independent practices, acquired.
Exit tracker PROPOSED
Who buys the asset five years later.
Corporate structure library PROPOSED
Publicly documented professional-corporation architectures, read against the corporate-practice doctrine.
Management agreement clauses PROPOSED
Control terms exposed in public litigation.
Bankruptcy aftermath PROPOSED
Wages, continuity of care, contracts, asset sales, ownership, licences and closures after a filing.
Medication access journey PROPOSED
Prescription, pharmacy, formulary, authorisation, benefit manager, denial, appeal.
Why this costs so much PROPOSED
Imaging, emergency visit, insulin, ambulance, admission, specialty drug, laboratory — each payment layer named.
Middlemen atlas PROPOSED
Benefit managers, administrators, purchasing organisations, clearinghouses and utilisation vendors.
Facilities, closures and delegated care
Where services stop existing, and who is responsible when the state contracts them out.
Service-line closure atlas PROPOSED
Obstetrics, behavioural health, trauma, paediatrics, dialysis, oncology.
Maternity desert dynamics PROPOSED
Why obstetric units close while demand persists.
Emergency boarding dashboard PROPOSED
Where state data allow the measurement.
Response and diversion PROPOSED
Emergency response and hospital diversion as local policy.
Correctional contractor database PROPOSED
Contracts, jurisdictions, ownership, litigation, findings, accreditation and enforcement history.
Deaths in custody registry PROPOSED
Structured and source-driven. Causation never inferred from sequence.
Jail healthcare contract library PROPOSED
Staffing requirements, penalties, indemnification, quality metrics and mortality review.
Delegated clinical function PROPOSED
What happens when the state contracts out clinical judgement but keeps constitutional responsibility.
The training pipeline and physician immigration
Four datasets the institute has already named as missing.
Residency visa transparency database PROPOSED
Programme, specialty, sponsorship, examination and graduation cutoffs, experience requirements, with a source date on every field.
Programme transparency index PROPOSED
Not a quality ranking. A disclosure ranking.
Opportunity database PROPOSED
Verified electives, observerships and research positions, with eligibility, fees and sponsorship. No paid listings.
Application waste calculator PROPOSED
Applications submitted to programmes whose own published screen excludes the applicant.
Filtering observatory PROPOSED
Which filters remove an applicant before any holistic review occurs.
Training money map PROPOSED
Who finances residency, and through which appropriation.
Slot expansion tracker PROPOSED
Which states and institutions are actually creating positions.
Visa workforce map PROPOSED
Waiver physicians by geography, where public data permit.
Alternative licensure tracker PROPOSED
Fifty-state legislation and implementation for internationally trained physicians.
Machine decisions, devices and evidence integrity
Accountability for systems that produce clinical text and coverage outcomes.
Incident registry PROPOSED
Documented healthcare failures: bias, fabricated text, erroneous denial, diagnostic error, security, model drift. Documented cases only.
Clinical provenance project PROPOSED
Who generated the text, who edited it, who adopted it, and who bears the consequence.
Scribe contract observatory PROPOSED
Vendor terms on data use, retention, training, privacy, indemnification and physician attestation.
The device boundary PROPOSED
Systems inside device regulation and systems outside it.
Algorithmic coverage registry PROPOSED
Public action and litigation involving automated coverage decisions.
Evidence quality scorecard PROPOSED
Prospective validation, external validation, trial evidence, post-market monitoring, subgroup analysis.
Device failure timeline PROPOSED
Clearance, signal, recall, litigation.
Warning-letter network PROPOSED
Company histories and recurring findings.
Trial noncompliance registry PROPOSED
Sponsor and institution patterns in results reporting.
Retraction tracker PROPOSED
High-impact clinical retractions and their guideline consequences.
Evidence lag observatory PROPOSED
Evidence to guideline to coverage to practice, measured in years.
Guideline politics PROPOSED
Conflicts, panel composition, evidence quality and strength of recommendation.
Consensus evolution PROPOSED
Where clinical consensus reversed, and what the reversal cost.
What medicine got wrong PROPOSED
Abandoned procedures, withdrawn medicines, reversed screening, and the policy that followed each.
Safety, harm and administrative injury
Harm that arrives through paperwork rather than through medicine.
Regulatory failure case studies PROPOSED
Signal, missed opportunity, harm, investigation, reform.
Never-event transparency PROPOSED
Where data permit publication.
Diagnostic delay observatory PROPOSED
System failure rather than individual fault.
Referral failure observatory PROPOSED
Where coordination structurally breaks.
Medical economics for physicians PROPOSED
Relative value units, conversion factors, capitation, shared savings, risk adjustment, facility fees and payment groupings, explained.
Productivity surveillance PROPOSED
How employers measure physicians, and what the measurement changes.
Metric gaming observatory PROPOSED
Documented incentives to improve the measure rather than the care.
Small practice survival PROPOSED
Reimbursement, security costs, credentialing, authorisation, records, malpractice, staffing, rent, compliance.
Independence versus consolidation PROPOSED
The economic comparison, done honestly.
The cost of being a physician PROPOSED
Training, licensure, registration, certification, education, malpractice, credentialing, dues.
Patient time as cost PROPOSED
Travel, waiting, authorisation, referral delay, appeals, paperwork.
Administrative harm PROPOSED
Harm arising from delay, denial, paperwork, credentialing error, network error, automated decision and regulatory lag.
What public records can and cannot support
Empirical work on institutional output is only as honest as its account of what the record leaves out. Six limits apply to almost every dataset in this program, and each one has to be stated on the face of the dataset rather than discovered by a reader later.
The denominator is usually missing
Published decisions are the visible end of a funnel whose earlier stages are not published. Complaints closed without action, investigations that go nowhere and accusations withdrawn before hearing are largely invisible. A count of outcomes is not a rate of anything until the denominator is identified or the absence of one is declared. This is the single most common defect in published enforcement analysis, and it has its own analysis on this site.
Settlement makes the reasoning disappear
A stipulated resolution can produce an order without findings, without evidence and without an explanation of why the sanction landed where it did. Datasets that read only reasoning-bearing documents will over-represent contested matters, which are systematically different from settled ones.
Selection runs in both directions
The matters that reach a hearing are selected by the parties’ own assessments of risk. A physician with strong evidence may be more willing to contest; an agency with a strong case may be less willing to settle. Any association between contesting and outcome is therefore contaminated by the decision to contest.
Coding is interpretation
“Quality of care”, “dishonesty” and “boundary violation” are categories imposed on documents that were not written to be coded. Every frame in this program publishes its coding rules, and every ambiguous record is recorded as ambiguous rather than forced.
Absence of precedent is not absence of pattern
Administrative decisions are largely non-precedential. Patterns still form, but they form as institutional custom, which means a pattern is a description of behaviour and never an entitlement a later respondent can claim.
Correlation will be reported as correlation
Where a variable travels with an outcome, the dataset says so in those terms. This program will not report an association as a cause, and will not describe a departure from a guideline as favouritism without evidence of the mechanism. Enforcement data need context is the standing statement of that discipline.
None of this is a reason not to do the work. It is the reason the work has to be done carefully rather than quickly, and it is why an item stays a question on this page until its record set can carry an answer. A dataset that cannot state its own limits is not evidence; it is decoration with numbers on it.
Sequence
The order is not arbitrary. Four criteria decide what gets built first: whether the primary records exist in publishable form today, whether the question has a reader who is currently unserved, whether the dataset compounds — that is, whether later work can be built on top of it — and whether the subject is one where this institute’s existing published corpus already supplies the doctrinal groundwork.
The discipline observatory
Records exist, are public, and are almost unusable in their published form. The doctrinal groundwork is already published here, so the dataset can be read against it immediately.
The training pipeline datasets
Four datasets this institute has already named as missing. The readership is specific, large and currently dependent on rumour.
Decision-rights and cascade maps
These compound: once built, they organise every later vertical, because most of the remaining questions are questions about who decides and what follows.
Everything measured across systems
The fifty-state atlases, the cross-profession comparisons and the influence trackers need the single-jurisdiction work to exist first, or there is nothing to compare.
Items outside that sequence are not dormant. Several will move early because a development forces them — a court decision, a rule, a closure, an enforcement action — and this institute’s standing rule is that a new development updates the existing record it affects rather than generating a fresh page beside it.
The point of view
A research program needs a stated position, or its choices of subject look arbitrary. Ten commitments, none of them partisan, all of them testable against this institute’s own output.
01
Evidence before institutional authority
An agency statement is evidence of the agency’s position. It is not proof that the position is correct.
02
Patient protection and physician due process are not opposites
A competent regulator does both. A regulator that trades one for the other has failed at both.
03
Clinical judgement should remain clinically accountable
Corporate, payer and administrative systems should not control clinical decisions invisibly.
04
Power should be traceable
Who decided? Under what authority? On what evidence? Subject to what review?
05
Allegation is not finding
The distinction is structural, not rhetorical, and it is coded into every dataset here.
06
Regulatory consequence should correspond to prospective risk
Where a sanction is justified by public protection, the risk it prevents should be identifiable.
07
Rules are judged by their operation
Not by their stated intent, and not by the intent of the people who wrote them.
08
Medicine should be explainable
Patients and physicians are entitled to understand the machinery that decides for them.
09
Institutional concentration deserves scrutiny
Conclusions come from documented structure and measured effect, never from the label attached to an owner.
10
Reform must be specific enough to implement
A proposal that cannot be drafted as text is a complaint, not a reform.
How an item becomes a published dataset
Question
The item is written as a question that a public record can answer. If no public record can answer it, it stays a question.
Source location
The issuing body’s own published documents are located and dated. Secondary summaries are leads, never citations.
Coding frame
Fields are defined before collection, including the fields that distinguish allegation from finding.
Record-level verification
Every record is checked against the primary document. A record that cannot be verified is not published.
Publication with limits
The dataset is published with its own account of what it does not cover and what it cannot establish.
Standing revision
New records are appended and dispositions added as dated entries. Nothing is overwritten, and no URL disappears.
The published corpus this program builds on: the policy library · research products · methodology · living dossiers · the event timeline · the enforcement ledger · intelligence · the reform agenda · the IMG institute · how medicine works · forecasts · search