Katherine Ho, "The Welfare Effects of Restricted Hospital Choice in the US Medical Care Market", Journal of Applied Econometrics, Vol. 21, No. 7, 2006, pp. 1039-1079. The data used in this paper come from a number of proprietary sources. Details are listed below. Any researcher wishing to access the data for the purpose of replicating the analysis should contact the author at: Katherine Ho Department of Economics, Columbia University 1037 International Affairs Building 420 West 118th Street New York, NY 10027 USA Email: kh2214 [T] columbia.edu The data sources are listed below. The hospital demand model was estimated using patient-level data for 1997 and 1998 from the MarketScan Database, Thomson MedstatŪ, Ann Arbor, MI. I used the Inpatient Admissions and Inpatient Services files from the Commercial Claims and Encounters Database. Data on hospital characteristics for 1997, 1998 and 2001 come from the AHA Annual Survey Database. The data are held by the American Hospital Association, Health Forum LLC, Chicago, IL. The population of consumers in each market by age, gender and zip code and median income by zip code were extracted from the "CensusCD" provided by GeoLytics, Inc., East Brunswick, NJ, 2003. Health insurer market shares and characteristics (premiums earned, number of enrollees, HMO/POS indicators and tax status) were taken from two datasets from Atlantic Information Services: "The HMO Enrollment Report, January 2003" and "HMO Directory 2003". Data can be purchased from http://www.aishealth.com. In cases where Atlantic Information Services data were not available data on premiums were taken from the Weiss Ratings' Guide to HMOs and Health Insurers, Fall 2002. I also used the plan age and number of physicians reported by this source. Information on purchasing these data is available at http://www.weissratings.com/products_hmo.asp. Health plan quality measures were taken from the Health Employer Data and Information Set (HEDIS) and the Consumer Assessment of Health Plans (CAHPS) 2000 data, both of which are published by the National Committee for Quality Assurance. Information is at the following website: http://www.ncqa.org/Communications/Publications/index.htm. I list below the variables utilized for each dataset. 1. Medstat MarketScan Database, 1997 and 1998 (a) Inpatient Admissions Table EMPZIP Postal ZIP code of the primary beneficiary's residence HOSPZIP Postal ZIP code of the hospital where the patient was admitted MSA Metropolitan Statistical Area of primary beneficiary PDX Principal diagnosis explaining the main reason for an admission PLANTYP Type of benefit plan. Sample restricted to 1 (Basic/major medical), 2 (Comprehensive) and 6 (PPO) UNIHOSP Unique hospital identifier YEAR Year during which service was rendered (b) Inpatient Services Table STDPLAC Place of service STDSVC Type of inpatient service Following advice from Medstat, encounters were defined as emergencies if STDPLAC = 23 or if STDPLAC=25 and STDSVC=77. 2. American Hospital Association Annual Survey Database, 1997, 1998 and 2001 ID AHA Identification Number REG AHA Region Code STCD AHA State Code HOSPNO AHA Hospital Number CNTRL Control Code (sample was restricted to hospitals with CNTRL<40) SERV Service Code (sample restricted to hospitals with SERV=10) OBHOS Obstetric care - hospital NICHOS Neonatal intensive care - hospital BRNHOS Burn care - hospital ALCHHOS Alcohol/drug abuse inpatient care - hospital ANGIOHOS Angioplasty - hospital BROOMHOS Birthing room - hospital MAMMSHOS Breast cancer screening/mammograms - hospital CCLABHOS Cardiac catheterization laboratory - hospital EMDEPHOS Emergency department - hospital ONCOLHOS Oncology services - hospital OHSRGHOS Open-heart surgery - hospital RADTHHOS Radiation therapy - hospital CTSCNHOS Computed-tomography (CT) scanner - hospital MRIHOS Magnetic resonance imaging (MRI) - hospital PETHOS Positron emission tomography (PET) - hospital SPECTHOS Single photon emission computerized tomography (SPECT) - hospital BDTOT Total facility beds set up and staffed MNAME Hospital name MLOCZIP Hospital ZIP code FTEMD Full time equivalent physicians and dentists FTERN Full time equivalent registered nurses MSACD Metropolitan statistical area code (MSA) MAPP1 Accreditation by Joint Commission on Accreditation of Health Care Organizations (JCAHO) MAPP2 Cancer program approved by American College of Surgeons MAPP3 Residency training approval by Accreditation Council for Graduate Medical Education MAPP5 Medical school affiliation reported to American Medical Association MAPP8 Member of Council of Teaching Hospitals of the Association of American Medical Colleges (COTH) MSANAME Hospital, MSA name 3. GeoLytics Census CD AreaKey Geography Key Code AreaName Area Name P077001 Median family income P008001 Total population P008002-P008040 Male population by age group P008041-P008079 Female population by age group 4. Atlantic Information Services Data, 2003 (a) The HMO Enrollment Report, January 2003 MSA-Specific enrollment for each HMO and POS product listed in the data excluding Medicaid-only and Medicare-only products. Where MSA-specific enrollment was not available, State or National enrollment was allocated to the MSA level using the methodology described in Appendix A. (b) The HMO Directory 2003 HMO Name Name of the plan Private/Public Private or public organization Tax Status For-profit or not-for-profit organization Premiums Earned Total premiums earned during the year Enrollment Number of enrollees (used to create premiums per member month) 5. Weiss Ratings Guide to HMOs and Health Insurers, January 2002 WEISSID Weiss insurer identification number COMMENCED Year commenced business PREMTOT01 Total premiums earned during the year PHYSEOY01 Total number of physicians at end of year 2001 ENROL01 Total number of enrollees at end of year 2001 National Committee for Quality Assurance Data, 2000 Health Employer Data and Information Set (HEDIS) ORG_ID Organization ID SUB_ID Submission ID PLANNAME NCQA plan name BRCANSC Breast cancer screening CERCANSC Cervical cancer screening CH_DEL Check-ups after delivery EYEEX Diabetic eye exams AD_IMM1 Adolescent immunization 1 AD_IMM2 Adolescent immunization 2 ADSMOKE Advice on smoking MI_30 Mental illness check-ups (b) Consumer Assessment of Health Plans (CAHPS) ORG_ID Organization ID SUB_ID Submission ID PLANNAME NCQA plan name CARE_QUICKLY Getting care quickly CARE_NEEDED Getting needed care The CAHPS and HEDIS variables are defined in Table 5 of the paper.