CMS claims data - sporedata/researchdesigneR GitHub Wiki

The Centers for Medicare & Medicaid Services (CMS) offers health coverage to over 100 million people through Medicaid, Medicare, the Children's Health Insurance Program (CHIP), and the Health Insurance Marketplace. Medicare is the US health insurance program for older people and those who qualify for Social Security Administration disability benefits. CMS data are population-based and can be linked to [NCHS] population health surveys to expand both data sources' analytic potential.

There are four types of Medicare (Plans):

• Part A (Hospital Insurance) covers home health care, hospice, lab tests, inpatient hospital care, skilled nursing facility (SNF), and surgery [1].

• Part B (Medical Insurance), which covers medically-necessary services like physicians' services and tests (diagnostic tests, injections, and procedures), outpatient care, durable medical equipment (oxygen tanks and wheelchairs), home health services, some preventive and other medical services [1].

Most Medicare beneficiaries are enrolled in Part A and/or B coverage. Together, Parts A and B comprise Original Medicare, which is a fee-for-service health plan.

• Part C or Medicare Advantage Plan offered by Medicare-approved private companies. The Medicare Advantage Plans include Health Maintenance Organizations (HMOs), Medicare Medical Savings Account Plans, Preferred Provider Organizations (PPO), Private Fee-for-Service Plans, and Special Needs Plans. These "bundled" plans include Medicare Parts A, B, and usually D. As of 2016, approximately 31% of Medicare beneficiaries were enrolled in Part C [1].

Because Medicare Plam C serves people who need additional coverage, it typically presents lower out-of-pocket costs than Original Medicare. It may offer extra benefits like dental, hearing, and vision services, Medicare prescription drug coverage, Medicare Supplement Insurance (Medigap), and outpatient coverage (durable medical equipment - wheelchairs and home oxygen equipment, emergency ambulance transportation, emergency room care, and laboratory testing - blood tests and urinalysis).

• Part D (Prescription Drug Coverage). Also known as Medicare Rx, Part D assists Medicare beneficiaries in paying for outpatient prescription drugs purchased at retail cost, mail order, home infusion, and long-term care pharmacies. As of 2014, approximately 70% of Medicare beneficiaries were enrolled in the Part D plan [1].

• Number and position of diagnosis code fields

  • High specificity is attained by restricting entry criteria based on ICD-9/10 diagnosis code(s) to the primary diagnosis field. The diagnosis code in the primary position does not always represent why a patient was admitted. Rather, it may be a condition with a relatively high reimbursement rate. As a result, when the diagnosis is present in, say, any of the first five positions on the claim, a research design might be altered by defining the condition of interest as being present [1].
  • Although the expansion of the number of ICD-9/10 diagnosis and procedure code fields on a claim from 9 and 6 to 25 did not noticeably affect the resulting data in the SAFs until 2011, it could result in an artificial increase in the estimates of disease burden, mainly if more sensitive definitions of disease are used [1].

• Completeness of inpatient claims at the end of each calendar year

  • Approximately two percent (2%) of claims for inpatient services at the end of each calendar year are omitted from the standard analytic files (SAF) provided to researchers. This can be explained by the fact that CMS extracts inpatient claims based on the discharge date, and a percentage of patients admitted into the hospital in one calendar year are only discharged the following year [1].
  • This omission can be problematic when studying major clinical events resulting in lengthy hospitalizations, such as serious infections or stroke. Culminated, these claims, together with those that are reprocessed or processed late, can affect up to 20% of December hospital admissions, resulting in spuriously low event rates at the end of a calendar year [1].

• The influence of CMS policy on data collection and coding

  • CMS reimbursement policies may affect how hospitals and health care providers code the provision of health care. As such, researchers working with Medicare data should stay abreast of CMS policy decisions and consider how these decisions may impact trends in disease and treatment utilization patterns in the data [1].
  • The Diagnosis-Related Group (DRG) determines how much CMS reimburses the hospital for the inpatient stay. The DRG system helps manage the costs of hospitalizations and can also influence the coding of disease within hospitalizations - DRG is assigned based on the principal diagnosis and procedure codes. For example, acute myocardial infarction (AMI) may be selected as the principal diagnosis instead of another disease condition because the DRG for AMI has a substantially higher reimbursement payment. Patients classified per a specific DRG are expected to use the same amount of hospital resources despite the intensity of the actual service provided [1]. See The 2019 Annual DRG Description Report
  • As part of the Affordable Care Act (ACA), the Hospital Readmissions Reduction Program was implemented by CMS and aims at reducing costs associated with hospital readmissions within 30 days after discharge by penalizing hospitals with higher than expected 30-day readmission rates for many selected conditions [1].
  • Despite the drop in hospital readmission rates since the program's implementation, there are concerns that the increased use of observation unit stays may still increase hospital readmission rates [1].

CMS claims data can be used to:

• Describe patterns of morbidity [2] and mortality [3] - Time to Event Models
• Describe the burden of disease [4], [5] - Exploratory Analyses , Time Series , Spatial Analyses
• Compare the effectiveness of pharmacologic therapies [6], [7], [8], [9] - Propensity Scores, Randomized Trial
• Evaluate the effects of provider practices on the delivery of care and patient outcomes - Mixed Models
• Examine the cost of care [10], [11], [12], [13], [14] - Cost-effectiveness, and
• Explore the effects of important Medicare policy changes on physician practices and patient outcomes [15], [16], [17], [18], [19], [20], [21], [22], [23] - Implementation Science.
• Can be linked to patient-level datasets using geographic information (zip codes, ZCTA, longitude/latitude.) and dates.

• Carrier - Carrier claims are non-institutional claims, however, this does not mean that they are outpatient claims.
• Durable Medical Equipment (DME) - contains final action claims data submitted to Durable Medical Equipment Regional Carriers (DMERCs). Some of the information contained in this file includes:
  • ICD-9 or ICD-10 diagnosis;
  • Services provided (HCFA Common Procedure Coding System (HCPCS) codes);
  • Dates of service;
  • Reimbursement amount;
  • DME provider number;
  • Beneficiary demographic information.
• Home Health Agency (HHA) - HHA has data for home health services. The information contained in this file includes:
  • Number and date of visits, type of visit (skilled-nursing care, home health aides, physical therapy, speech therapy, occupational therapy, and medical social services);
  • ICD-9 or ICD10 diagnosis;
  • HHA provider number
  • Beneficiary demographic information. An HHA bill may cover services provided over a period of time, not on a single day. Because the claim total payment amount is repeated on every record associated with a claim, sort the file on patient_id, year, claim_id, and rec_count. It keeps the first record in the sort (if first claim_id).
• Hospice - The Hospice contains claims data submitted by Hospice providers. The information contained in this data includes:
  • The level of hospice care received (e.g., routine home care, inpatient care);
  • Terminal diagnosis (ICD-9 or ICD-10 diagnosis);
  • The dates of service;
  • Hospice provider number;
  • Beneficiary demographic information.
  • Inpatient;
• Outpatient - The outpatient file contains data from institutional outpatient providers (hospital outpatient departments, rural health clinics, renal dialysis facilities, outpatient rehabilitation facilities, comprehensive outpatient rehabilitation facilities, community mental health centers). The information contained includes:
  • diagnosis and procedure codes;
  • dates of service;
  • Reimbursement amounts,
  • Facility provider number;
  • Revenue center codes and beneficiary demographic information.
  • ICD-9 or ICD-10 procedure codes, the reporting of these codes since 2000, services from the outpatient bill have been captured from CPT/HCPCS codes and revenue centers. Definitions for revenue center codes may be obtained by contacting ResDAC or CMS directly.
• Skilled Nursing Facility - SNF is included in the MEDPAR file. MEDPAR contains one summarized record per admission. Each record includes:
  • Up to 25 ICD-9 diagnoses and 25 ICD-9 procedures provided during the hospitalization.

More details are available here.

• CMS data are collected for the purpose of making healthcare payments, and not for research, there are limitations to the data that you should consider when analyzing;
• Only includes a diagnosis documented via the International Classification of Diseases, Ninth Revision (ICD-9) or ICD-10codes, and its difficult to assess surgical complications;
• There is no physiological or biochemical patient information, such as vital signs, laboratory test results, and pathology results;
• There are no timestamps during a hospital stay. This can limit the study of how care progresses or when events and complications occur during hospitalization;
• Lack of data on uncovered services or benefits and managed care enrollee information. This limits the evaluation of outpatient utilization patterns;
• The data are derived from billing data, allowing inconsistent documentation for comorbidity and severity;
• Researchers using the Home Health Agency data need to be aware of changes over time in how CMS codes HHA services;
• If the stay is long, there may be more than one claim per stay, resulting in more than a single MEDPAR record. This occurs most frequently for stays in SNFs as these often span several months;
• SNFs records often have no discharge date as persons remain in institutions beyond the period of Medicare coverage.

• Practical Guide to Surgical Data Sets: Medicare Claims Data
• Use of the Medicare database in epidemiologic and health services research: a valuable source of real-world evidence on the older and disabled populations in the US
• Identifying Cancer-Directed Surgeries in Medicare Claims: A Validation Study Using SEER-Medicare Data

• Medicare claims data dictionary

[1] Mues KE, Liede A, Liu J, Wetmore JB, Zaha R, Bradbury BD, Collins AJ, Gilbertson DT Use of the Medicare database in epidemiologic and health services research: a valuable source of real-world evidence on the older and disabled populations in the US. Clinical epidemiology. 2017;9:267.

[2] Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. New England Journal of Medicine. 2009 Apr 2;360(14):1418-28.

[3] Gornick ME, Eggers PW, Reilly TW, Mentnech RM, Fitterman LK, Kucken LE, Vladeck BC. Effects of race and income on mortality and use of services among Medicare beneficiaries. New England journal of medicine. 1996 Sep 12;335(11):791-9.

[4] Schermerhorn ML, Buck DB, O’Malley AJ, Curran T, McCallum JC, Darling J, Landon BE. Long-term outcomes of abdominal aortic aneurysm in the Medicare population. New england journal of medicine. 2015 Jul 23;373(4):328-38.

[5] Wetmore JB, Peng Y, Jackson S, Matlon TJ, Collins AJ, Gilbertson DT. Patient characteristics, disease burden, and medication use in stage 4-5 chronic kidney disease patients. Clinical nephrology. 2016 Feb 1;85(2):101-11.

[6] Schneeweiss S, Solomon DH, Wang PS, Rassen J, Brookhart MA. Simultaneous assessment of short‐term gastrointestinal benefits and cardiovascular risks of selective cyclooxygenase 2 inhibitors and nonselective nonsteroidal antiinflammatory drugs: an instrumental variable analysis. Arthritis & Rheumatism: Official Journal of the American College of Rheumatology. 2006 Nov;54(11):3390-8.

[7] Wang PS, Schneeweiss S, Avorn J, Fischer MA, Mogun H, Solomon DH, Brookhart MA. Risk of death in elderly users of conventional vs. atypical antipsychotic medications. New England Journal of Medicine. 2005 Dec 1;353(22):2335-41.

[8] Izurieta HS, Thadani N, Shay DK, Lu Y, Maurer A, Foppa IM, Franks R, Pratt D, Forshee RA, MaCurdy T, Worrall C. Comparative effectiveness of high-dose versus standard-dose influenza vaccines in US residents aged 65 years and older from 2012 to 2013 using Medicare data: a retrospective cohort analysis. The Lancet Infectious diseases. 2015 Mar 1;15(3):293-300.

[9] Yun H, Delzell E, Saag KG, Kilgore ML, Morrisey MA, Muntner P, Matthews R, Guo L, Wright N, Smith W, Colón-Emeric C. Fractures and mortality in relation to different osteoporosis treatments. Clinical and experimental rheumatology. 2015 May;33(3):302.

[10] Gilden DM, Kubisiak JM, Pohl GM, Ball DE, Gilden DE, John WJ, Wetmore S, Winfree KB. Treatment patterns and costeffectiveness of first line treatment of advanced non-squamous non-small cell lung cancer in Medicare patients. Journal of Medical Economics. 2017 Feb 1;20(2):151-61.

[11] Gozalo P, Plotzke M, Mor V, Miller SC, Teno JM. Changes in Medicare costs with the growth of hospice care in nursing homes. New England Journal of Medicine. 2015 May 7;372(19):1823-31.

[12] Padegimas EM, Verma K, Zmistowski B, Rothman RH, Purtill JJ, Howley M. Medicare reimbursement for total joint arthroplasty: the driving forces. JBJS. 2016 Jun 15;98(12):1007-13.

[13] Lairson DR, Parikh RC, Cormier JN, Chan W, Du XL. Cost–Utility Analysis of Chemotherapy Regimens in Elderly Patients with Stage III Colon Cancer. Pharmacoeconomics. 2014 Oct;32(10):1005-13.

[14] Boero IJ, Gillespie EF, Hou J, Paravati AJ, Kim E, Einck JP, Yashar C, Mell LK, Murphy JD. The impact of radiation oncologists on the early adoption of hypofractionated radiation therapy for early-stage breast cancer. International Journal of Radiation Oncology, Biology, Physics. 2017 Mar 1;97(3):571-80.

[15] O'Malley AS, Reschovsky JD, Saiontz-Martinez C. Interspecialty communication supported by health information technology associated with lower hospitalization rates for ambulatory care–sensitive conditions. The Journal of the American Board of Family Medicine. 2015 May 1;28(3):404-17.

[16] Ko CW, Dominitz JA, Green P, Kreuter W, Baldwin LM. Specialty differences in polyp detection, removal, and biopsy during colonoscopy. The American journal of medicine. 2010 Jun 1;123(6):528-35.

[17] Hirth RA, Turenne MN, Wheeler JR, Pan Q, Ma Y, Messana JM. Provider Monitoring and Pay‐for‐Performance When Multiple Providers Affect Outcomes: An Application to Renal Dialysis. Health Services Research. 2009 Oct;44(5p1):1585-602.

[18] Arneson TJ, Li S, Gilbertson DT, Bridges KR, Acquavella JF, Collins AJ. Impact of Centers for Medicare & Medicaid Services national coverage determination on erythropoiesis‐stimulating agent and transfusion use in chemotherapy‐treated cancer patients.. Pharmacoepidemiology and drug safety. 2012 Aug;21(8):857-64.

[19] Zuckerman RB, Sheingold SH, Orav EJ, Ruhter J, Epstein AM. Readmissions, Observation, and the Hospital Readmissions Reduction Program. New England Journal of Medicine. 2016 Apr 21;374(16):1543-51.

[20] Barr CD, Diez DM, Wang Y, Dominici F, Samet JM. Comprehensive smoking bans and acute myocardial infarction among Medicare enrollees in 387 US counties: 1999-2008. American journal of epidemiology. 2012 Oct 1;176(7):642-8.

[21] Shahinian VB, Kuo YF, Gilbert SM. Reimbursement policy and androgen-deprivation therapy for prostate cancer. New England Journal of Medicine. 2010 Nov 4;363(19):1822-32.

[22] Chertow GM, Liu J, Monda KL, Gilbertson DT, Brookhart MA, Beaubrun AC, Winkelmayer WC, Pollock A, Herzog CA, Ashfaq A, Sturmer T. Epoetin alfa and outcomes in dialysis amid regulatory and payment reform. Journal of the American Society of Nephrology. 2016 Oct 1;27(10):3129-38.

[23] Wang C, Kane R, Levenson M, Kelman J, Wernecke M, Lee JY, Kozlowski S, Dekmezian C, Zhang Z, Thompson A, Smith K. Association between changes in CMS reimbursement policy and drug labels for erythrocyte-stimulating agents with outcomes for older patients undergoing hemodialysis covered by fee-for-service Medicare. JAMA internal medicine. 2016 Dec 1;176(12):1818-25.