Chapter 36 — Exercises

How to use these. Section D is arithmetic and should be drilled until building and checking a risk score is automatic. Section G is the reasoning section — MEAT applied to real-looking documentation — and it is where risk-adjustment coders are actually made; do those slowly, condition by condition, and write down which of the four criteria each condition met. Section H has no arithmetic in it at all and is the section most likely to matter to your career. Items marked † have worked solutions in the answers appendix. No answers appear in this file.

Every coefficient, benchmark, score, and dollar figure in these exercises is a constructed teaching figure. Risk models, their category definitions, their code mappings, and every coefficient are revised annually, and the model version in force differs by contract year and by program. Verify current values with the plan or at CMS. ICD-10-CM changes every October 1; CPT changes every January 1; HCPCS Level II changes quarterly.

Section A — Paying for a population instead of a visit (items 1–4)

A.1 In one sentence each: what triggers a payment under fee-for-service, and what triggers a payment under risk-adjusted capitation?

A.2 † A diagnosis code is omitted from an office claim. Describe, step by step, what the practice would see happen — under fee-for-service, and then under a risk-adjusted arrangement. Your answer to the second half should be short, and the reason it is short is the point.

A.3 Name the two jobs a diagnosis code does under a risk-adjusted arrangement, and say who reads it for each job and when.

A.4 A practice manager returns from a meeting with a value-based contract and asks the coder, "does this change what you do?" Write the coder's answer in under 120 words. It must contain both halves: what does not change about the coding, and what does change about the work.

Section B — What risk adjustment is trying to fix (items 5–8)

B.5 † Chapter 2's Case Study 2 lists "the rate was set without adequate risk adjustment" as the first documented cause of 1990s capitation failures. Explain the mechanism in your own words, then state what an unadjusted per-member payment pays an organization to do about sick patients.

B.6 "The model predicts cost, not sickness." Give one consequence of that sentence that would surprise a clinician, and one that would surprise a coder.

B.7 A risk score of 1.0 means what, exactly? A score of 1.6? Why is the model normalized at all?

B.8 Explain why under-reporting and over-reporting are the same error with opposite signs in this setting. Reference Chapter 5 §5.8 and name three separate things under-reporting damages besides the payment.

Section C — Hierarchical condition categories (items 9–12)

C.9 Define a hierarchical condition category in one sentence, then say what the word hierarchical is doing in the name.

C.10 † A coder new to risk adjustment reviews a chart with fourteen accurate ICD-10-CM codes and finds that only two of them map to a condition category. She concludes she has made a mistake. Explain why she has not, and what a risk model is actually selecting for.

C.11 Name four different risk-adjustment models in use in the United States, say which population each serves, and give one reason they legitimately disagree about which conditions carry weight.

C.12 Why does this book print no HCC numbers and no coefficients as fact? Give the two reasons, and then state the habit the policy is meant to build.

Section D — Building and checking a risk score (items 13–18)

All coefficients and benchmarks below are constructed teaching figures.

D.13 † Build the score. A member's record for the collection year supports the following, and all have been documented and reported on face-to-face encounters:

   demographic component (female, 70-74, community, non-dual)   0.346
   chronic obstructive pulmonary disease                        0.328
   diabetes with chronic complications                          0.302
   chronic kidney disease, moderate stage                       0.127
   congestive heart failure                                     0.331
   interaction: heart failure + diabetes                        0.121

(a) Compute the RAF. (b) At a benchmark of \$1,000.00 per member per month, compute the monthly and the annual payment. (c) Show every check.

D.14 Using D.13's figures, compute the score if the heart failure had never been documented. State the total loss and explain why it is larger than the heart failure coefficient by itself.

D.15 † Using D.13's figures, compute the score if the diabetes had been reported as without complications at a coefficient of 0.105 and the kidney disease had never been reported. State the difference from D.13 and, at the same benchmark, what it is worth for the year.

D.16 A colleague proposes reporting both the with-complications and the without-complications diabetes codes "so the model can pick." Compute what the model counts. Then explain, in two sentences, what the colleague has misunderstood about how a hierarchy works.

D.17 Two members have identical documented conditions. One has a demographic component of 0.223 and the other 0.512. Compute both scores using D.13's disease and interaction terms, and explain in one sentence why a chart cannot close that gap.

D.18 † A plan tells a practice that a particular member's score is 1.42 and the county benchmark is \$1,150.00 PMPM. (a) Compute the annual payment for that member. (b) The practice's administrator concludes the practice will receive that amount. Explain what is wrong with that conclusion, and name the document that decides what the practice receives.

Section E — Hierarchies, trumping, and interactions (items 19–22)

E.19 State the trumping rule in one sentence, then state the design consequence: what does the hierarchy protect against, and what does it deliberately not protect against?

E.20 † Explain why the asymmetry in a hierarchy — over-reporting gains nothing, under-reporting loses everything — makes accurate coding the obviously correct strategy rather than merely the ethical one. Then give the one situation in which a deleted code changes a score in the other direction.

E.21 A member has documented diabetes with chronic complications, chronic kidney disease stage 4, and chronic obstructive pulmonary disease. Which of these trump which? Answer from the structure of disease families rather than from any published table.

E.22 Why can an interaction term never be reported on a claim? What is the only way to earn one?

Section F — The annual reset (items 23–26)

F.23 State the annual reset rule, including which year's diagnoses pay for which year.

F.24 † A physician objects: "The amputation did not grow back. Why am I documenting it again?" Write the response you would actually give — one that concedes the point before it explains the rule. Under 150 words.

F.25 For each of the following, state whether it recaptures a chronic condition for the collection year, and why: (a) an entry on the reviewed problem list; (b) a laboratory result showing an eGFR of 48; (c) a medication refill authorized by telephone; (d) a face-to-face office visit at which the condition is assessed with a status and a plan; (e) last year's note, copied forward unchanged into this year's; (f) a claim for a venipuncture carrying the diagnosis.

F.26 † Account 10-4471's knee is established as M17.11 after the April imaging (Chapter 22). Osteoarthritis is chronic, permanent, and frequently disabling. Explain why it must still be documented every year, and why a coder should not assume it carries a risk weight. Then state the general principle about what risk models select for.

Section G — MEAT against documentation (items 27–31)

For each item, work condition by condition. For every condition, state which of the four MEAT criteria the note satisfies — if any — and what is reportable. All notes are constructed teaching examples.

G.27 † "Assessment/Plan: 1. Type 2 diabetes mellitus — stable, continue metformin, A1c ordered today. 2. Essential hypertension — at goal, continue lisinopril, no change. 3. Hyperlipidemia — continue atorvastatin, lipid panel ordered. 4. Right knee pain — new complaint this visit, consistent with a degenerative process; no definitive diagnosis established today. Problem list reviewed: type 2 diabetes mellitus, essential hypertension, hyperlipidemia, chronic kidney disease stage 3a, obesity."

Work all five. State exactly which one fails and on how many criteria.

G.28 "HPI: 62-year-old with COPD on home oxygen, here for follow-up. Reports two exacerbations since last visit. Also has heart failure and takes furosemide. Assessment/Plan: COPD — increase inhaled corticosteroid, pulmonary referral placed. Continue current medications."

Two conditions are clinically present. How many are reportable, and what specifically is wrong with the documentation of the other?

G.29 † "Assessment: 1. Chronic systolic heart failure — weight up 4 lb, mild ankle edema. Increase furosemide to 40 mg daily, recheck BMP in one week, call if weight rises another 3 lb. 2. History of CVA — no residual deficits, aspirin continued. 3. History of CHF."

Item 3 is a trap on two separate grounds. Name both. (One is about redundancy; one is about a phrase this book has warned you about since Chapter 12.)

G.30 Explain why "stable" is fully reportable, using Chapter 15 §15.5's definition of the word. Then explain the converse trap: why a patient described as stable is not thereby a low-risk member.

G.31 † A practice's template auto-populates the assessment with the phrase "Chronic conditions reviewed and stable; medications refilled." The physicians like it and it saves time. Write the memorandum you would send the practice manager: what the phrase supports, what it does not support, what it does to a risk score, what it does to an evaluation and management level (Chapter 15 §15.10), and what you propose instead. Do not exceed 300 words.

Section H — Chart review, compliance, and detection (items 32–35)

H.32 Define prospective and retrospective chart review, then state the single design feature that separates a legitimate prospective program from an improper one.

H.33 † A vendor proposes a retrospective review program on a contingency arrangement — a percentage of the incremental revenue identified. Write the three questions you would ask, in order, and state what each answer would tell you. Then state whether a contingency arrangement is per se improper, and defend the answer.

H.34 Name five ways an unsupported risk-adjustment submission is detected. For each, say roughly who finds it and whether a medical record is required to find it.

H.35 † Your practice's own review finds that nine diagnoses submitted for last year's collection period are not supported by the records. List, in order, what happens next — including the clock that starts, the chapter and section of this book that owns it, and who must make any change to a clinical record.

Section I — Quality measures and value-based contracts (items 36–39)

I.36 Classify each as a structure, process, or outcome measure, and say what each one cannot tell you on its own: (a) the practice maintains a patient registry; (b) the percentage of diabetic patients who had a retinal exam this year; (c) the percentage of diabetic patients whose most recent A1c was below 8.0%; (d) the thirty-day readmission rate.

I.37 † 0001F and 3044F are both CPT Category II codes and they report different families of measure. Say which is which and why. Then explain to a physician, in three sentences, why a coder spends time assigning codes that will be adjudicated at nothing.

I.38 Compute the shared-savings distribution. [constructed] Attributed members 4,000; benchmark \$11,500.00 per member per year; actual expenditures \$43,240,000.00; minimum savings rate 2.0%; sharing rate 50%; quality multiplier 85%. Show the benchmark total, the gross savings, the savings percentage, whether the minimum savings rate is met, and the distribution. Show every check.

I.39 † Name the four contract terms that decide whether a value-based arrangement can be met, and write the question you would ask the plan about each. Then explain why more complete diagnosis coding raises revenue in a capitated arrangement but usually does not manufacture savings in a shared-savings one.

Section J — The Encounter (item 40)

J.40 † The full revisit, in your own words. Using Account 10-4471's March 14 note and nothing else:

(a) Apply MEAT to all five conditions on the record and produce the grid.

(b) State the two questions involved in the diabetes code — the one Chapter 9 §9.7 answered and the one Chapter 36 answered — and explain in three sentences why both answers can be correct simultaneously.

(c) Write the query you would send. It must quote only the relevant excerpt, offer the negative answer with equal weight, ask about this encounter specifically, and name no code.

(d) State what happens to the code if the answer is yes, and what happens if the answer is no.

(e) Explain, in one sentence, why a query cannot fix this problem in general — and name the process that can.

(f) The patient is 58 and commercially insured. Explain why this chapter's risk arithmetic is a lens for this file rather than its money, and then make the strongest honest argument for why the finding matters to Northgate anyway.