Data Cart

Your data extract

0 variables
0 samples
View Cart
Jump to beginning of year (scroll up to move forward in time):
2023 2022 2021 2020 2019 2018

October 4, 2023

    New Data.

    • IPUMS MEPS now includes 2021 data; over 900 variables (both annual person-level variables and round-level variables) from the 2021 file are now available through the extract system. Variables new to this year include COVID variables (CVDBOOSTER and CVDVAX) and variables from the Social Determinants of Health (SDOH) supplement (see description of SATISFIED for full list of variables).

    Changed data.

    • The width of HHAGCHTOT has increased from 6 to 7 columns wide to accomodate larger values.

    DOI Note.

    • Version D071.V2.2 of IPUMS MEPS was skipped due to production delays related to changes made to accommodate the 2020 extended MEPS panel. See the IPUMS Health Surveys DOI project page for a complete listing of IPUMS MEPS DOI versions.

July 27, 2023

    Changed data.

    • The width of STRATANN has increased from 4 to 8 columns wide to accommodate the addition of YEAR as a prefix. Efforts to reproduce MEPS published estimates on the AHRQ data tools website indicated that we were able to perfectly reproduce point estimates but only able to imperfectly reproduce standard errors when using the value of strata directly from the annual AHRQ files in data sets combining two or more years. By adding YEAR as a prefix to the value of strata from the annual files, we are able to perfectly reproduce published point estimates and standard errors from the AHRQ data tools website. For more information, see the IPUMS MEPS note on Analysis and Variance Estimation with MEPS.

    New Data.

    • IPUMS MEPS now includes records derived from the Prescribed Medicines file. These include over 70 new variables across three new record types comprised of Medication-Round Fill (F), Medication-Round (M), and Event-Medication-Round link (K) records collected on the MEPS since 1996. F records appear and can be added to extracts under the RX Medicines menu. M record variables are automatically added when an F record variable is selected and K record variables are automatically added when an Event (E) record variable is selected.

April 21, 2023

    New Data.

    • IPUMS MEPS now includes 2020 FYC data; over 900 variables (both annual person-level variables and round-level variables) from the 2020 file are now available through the extract system. Variables new to this year include COVID variables (CVDLAYCA, CVDLAYDN, CVDLAYPM) and a wage outlier flag OUTFLAG. The MEPS interview structure and sample design were adjusted to accommodate nonresponse bias due to the COVID-19 pandemic. Face-to-face interviews were replaced with telephone interviews and Panel 23 was extended to nine rounds (rather than the standard five), which led to changes in how weights were developed for FY20. For more information, see the MEPS Full Year 2020 user note.

    Modified Data.

    • EXPMCPAY and RXEXPMC were changed from 6 wide to 7 wide to accommodate for larger sum of payments.

June 9, 2022

May 2, 2022

    New Data.

    • IPUMS MEPS now includes records derived from the Medical Condition and Event files. These records include over 100 new variables collected on the MEPS since 1996. For more information, please see our overview of the event and medical conditions variables, including how those variables are offered through IPUMS MEPS, and general considerations to keep in mind when conducting analyses using these variables.

    Corrected data.

    • We corrected an error in the values of BEGRFMRD in 2015 and 2016. Previously BEGRFMRD contained the values of ENDRFMRD by mistake.

January 24, 2022

    Modified documentation.

    • IPUMS MEPS staff reviewed the set of variables from the Access to Care section of MEPS, finalized the variable documentation accordingly, and introduced the new variables YRSINUSC, USCOWNGRP, DELAYDTCST, and DELAYRXCST. The Access to Care section gathers information on usual source of care for all family members, characteristics of usual source of care health providers, and barriers family members have faced in obtaining needed health care. Please see ACCESSELIG for an overview of the available variables.

    Corrected data.

    • Previously, for the 1996 sample, USHOSYNONE was incorrectly based on the original MEPS variable that NOUSLYNONE is based on. Additionally, DIFCARFMYEQP, DIFCARFMYWRK, and DIFCARFMYWHR were incorrectly based on the original MEPS variable that DIFCARFMYIND is based on. The 1996 sample is now using the correct original MEPS variables.
    • The codes for the "yes" and "no" values for the 1996 sample of USHOSYDK were accidentally flipped. The error has been corrected.
    • Previously, the codes for the "Unknown-not ascertained" and "Unknown-don't know" values for the 1996 sample of USCCHNG were flipped. The error has been corrected.

    Modified data.

    • The variable YRSINUS was renamed to YRSINUSG. Additionally, the labels for the "unknown" categories have been changed. They were previously "7 - Refused," "8 - Don't know", and "9 - Not ascertained." They are now "7 - Unknown-refused," "8 - Unknown-not ascertained," and "9 - Unknown-don't know."
    • The label for the value of 7 in WHYNOUSLPL was previously "No place available in area" and has been updated to "Usual source of care provider not available in area." Additionally, the code for "Other reason" was changed from 30 to 40. Previously, for 1996-2001, the codes for "Other insurance-related reason" was incorrectly mapped to "No health insurance." The error has been corrected. Finally, the code 12 was changed from "No health insurance" to "Usual source of care doctor is somewhere else." "No health insurance" now takes the code 32.
    • To accommodate greater detail in later samples, the codes for TYPUSCTRANS have been expanded from a width of 1 to 2. Previously, the codes were "0 - NIU," "1 - Drive/Is driven," "2 - Taxi, bus, train, other public transport," "3 - Walks," "4 - Other way," "7 - Unknown-refused," "8 - Unknown-not ascertained," and "9 - Unknown-don't know." Now, the codes are "0 - NIU," "10 - Drives or is driven," "11 - Drives," "12 - Is driven," "20 - Taxi, bus, train, other public transport," "30 - Walks," "40 - Other way," "97 - Unknown-refused," "98 - Unknown-not ascertained," and "99 - Unknown-don't know."
    • The values for USUALPLFMY were adjusted to accommodate the addition of a new code. The codes 1-7 and 97-99 remain unchanged. Previously, the codes were "8 - Seldom sick/don't need usual source," "9 - Other insurance-related reason," and "10-Other." Now, the codes are "8 - Cost-related reason," "9 - Seldom sick/don't need usual source," "10 - Other insurance-related reason," and "11 - Other."
    • Previously, the 2013 sample for USCPRLANG only included responses from individuals in Panel 18. It now includes the repsonses of those in both Panel 17 and Panel 18.

December 9, 2021

    New Data.

    • IPUMS MEPS now includes 75 new variables from the Child Preventive Health Supplement. For more information see LESSHLTHY.

    Modified Data.

    Corrected Data.

    • Previously there was a mistake in 2012 for ADHECR (Rating of all health care received: last 12 months). A rating of 1 was coded as 10, a rating of 2 was coded as 1, a rating of 3 was coded as 2, a rating of 4 was coded as 3, a rating of 5 was coded as 4, a rating of 6 was coded as 5, a rating of 7 was coded as 6, a rating of 8 was coded as 7, a rating of 9 was coded as 8, and a rating of 10 was coded as 9. This has been corrected so that a rating of 1 is coded as 1, a rating of 2 is coded as 2, etc.

October 26, 2021

    Modified data.

    • Previously, in hierarchical extracts, the values for the record types "person" and "round" appeared as "2" and "3," respectively. Now, the record types appear as "P" and "R."

September 14, 2021

    New data.

    • IPUMS MEPS now includes 2019 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.

    Corrected data.

    • The variable CPI2009 previously reported the incorrect CPI conversion factor for 2016-2018. Additionally, IPUMS NHIS staff updated the values of CPI2009 to use the most recently published statistics by the Federal Reserve Board. These changes are minimal and should have no substantive effect on analaysis. The errors have been corrected and modifications have been made both in the data and the corresponding user note.
    • Previously, for the 2018 MEPS sample DIAEYEXCALYR erroneously referred to the MEPS variable DSEY1753, which has a reference period of the previous calendar year. DIAEYEXCALYR now correctly refers to the original MEPS variable for having a dilated eye exam during the current calendar year.
    • Previously, for the 2018 MEPS sample DIAFTCKCALYR erroneously referred to the MEPS variable DSFT1753, which has a reference period of the previous calendar year. DIAFTCKCALYR now correctly refers to the original MEPS variable for having feet checked during the current calendar year.
    • Previously, for the 2018 MEPS sample DIACHLCALYR erroneously referred to the MEPS variable DSCH1753, which has a reference period of the previous calendar year. DIACHLCALYR now correctly refers to the original MEPS variable for having blood cholesterol checked during the current calendar year.
    • Previously, for the 2018 MEPS sample DIAFLCALYR erroneously referred to the MEPS variable DSFL1753, which has a reference period of the previous calendar year. DIAFLCALYR now correctly refers to the original MEPS variable for receiving a flu vaccination during the current calendar year.

    Modified data.

    • DEPRESS2WK has been renamed to PHQDEP to match the IPUMS NHIS names of a substantively similar variable that was introduced as part of the 2019 NHIS sample.
    • The codes for TYPPLSICK were changed and expanded to be three digits wide. The categories of TYPPLSICK were previously: 00 = "NIU"; 10 = "Clinic or health center"; 11 = "Health center"; 12 = "Public health clinic"; 13 = "Clinic, unspecified type"; 14 = "Community/immigrant/rural clinic"; 15 = "Company or school clinic/center"; 16 = "Urgent care/walk-in clinic"; 17 = "Migrant clinic"; 18 = "Rural health center"; 19 = "Community or family health center"; 20 = "Doctor's office or HMO"; 21 = "Doctor's office"; 22 = "HMO"; 23 = "Doctor's clinic"; 24 = "Private doctor's office in hospital"; 30 = "Hospital emergency room"; 40 = "Hospital outpatient department"; 41 = "County/city/public/county hospital outpatient"; 42 = "Private/other hospital outpatient clinic"; 50 = "Some other place"; 51 = "Psychiatric hospital/clinic"; 52 = "VA hospital/clinic"; 53 = "Military health care facility"; 54 = "Military or VA health care facility"; 55 = "At home"; 56 = "Non-hospital place"; 57 = "Other places (1985: Includes known HMOs)"; 60 = "Doesn't go to one place most often"; 97 = "Unknown-refused"; 98 = "Unknown-not ascertained"; and 99 = "Unknown-don't know."
    • Codes for EDUC have been updated to accommodate a change in the coding of the underlying education data introduced in the 2019 NHIS data, which collapsed grades 1-11 into a single category. The coding structure for MEPS is the same as previous years, but the changes to the NHIS coding structure required significant revising of existing categories. Previously, the categories of EDUC were: 100 = "Grade 8 or less"; 101 = Never attended/kindergarten only; 102 = Grade 1; 103 = Grade 2; 104 = Grade 3; 105 = Grade 4; 106 = Grade 5; 107 = Grade 6; 108 = Grade 7; 109 = Grade 8; 200 = Grade 9-12, no diploma; 201 = Grade 9; 202 = Grade 10; 203 = Grade 11; 204 = 12th grade, no diploma; 300 = High school diploma or GED; 301 = High school graduate; 302 = GED or equivalent; 400 = Some college, no 4 yr degree; 401 = Some college, no degree; 402 = AA degree: technical/vocational/occupational; 403 = AA degree: academic program; 500 = Bachelor's degree (BA,AB,BS,BBA); 600 = Master's, Professional, or Doctoral Degree; 601 = Master's degree (MA,MS,Med,MBA); 602 = Professional degree (MD,DDS,DVM,JD); 603 = Doctoral degree (PhD,EdD); 604 = Other degree; 996 = No degree, years of education unknown; 997 = Unknown--refused; 998 = Unknown -- not ascertained; 999 = Unknown -- don't know.

October 8, 2020

    New data.

    • IPUMS MEPS now includes 2018 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.

    Modified data.

    • ADAPPT, ADDPRS, and ADEZUN have been renamed to VISITYRNO, DEPRESS2WK, and EZINFOFREQ, respectively, to match the IPUMS NHIS names of substantively similar variables.
    • MEPSID and MEPSIDRD have taken on a new meaning. Beginning in 2018, AHRQ began offering a unique record identifier on the original MEPS files as a combination of PANEL + DUPERSID. IPUMS MEPS staff created this variable for the 1996-2017 samples and now offer this identifier as MEPSID and MEPSIDRD. MEPSID2 and MEPSID2RD now take on the old meaning of the IPUMS-created MEPSID as a combination of DUID + PID +PANEL, but these variables will not be offered from 2018 forward.
    • To accommodate the introduction of a new code in the original data, the code for "Other" was changed from 10 to 11 for WHYNOWK2RD.

    Corrected data.

    • On extracts containing round data structured as wide instead of long, the negative values of FTOTINCCPS, FTOTINCMEPS, INCTOT, FTOTVAL, INCBUS, INCPROP, INCTRST, POVLEV, and MCS were incorrectly written out as missing values on all data formats except CSV. The issue has been corrected and negative values are now being correctly handled.

June 16, 2020

    New data.

    • IPUMS MEPS now includes the full set of cancer diagnosis, age of diagnosis, and remission status variables. For more information, see CNBLAD. The variables AGELAST, indicating a respondent's age the last time they were eligible for MEPS, and HINOTCOV, an indicator for respondents with no health insurance, are also now available.

    Corrected data.

    • CPI2009 was updated to reflect the 2017 CPI U adjustment factor for the 2017 sample. Previously, all values were zero. Additionally, we corrected an error in the codes for the variable USCPRFAC. "Person" was erroneously coded as "facility" and "facility" coded as "person." This error has been corrected.

    Modified data.

    • The coding for the 1999 sample of ACCESSELIG was updated to increase comparability across years. The cases that were previously coded to "Not in universe" are now coded to "No."

    Corrected documentation.

    • The universe for USCPRFAC was incorrect. It previously indicated that persons without a usual source of care provider (USUALPL) were asked this question. The universe was updated to properly reflect that all persons who do have a usual source of care (USUALPL) were asked this question.
    • The universe for USCPRTYP has been updated to state that persons with a usual source of care provider (USUALPL) who is either a person, a person in a facility, or was not ascertained are in universe for USCPRTYP. The universe previously did not include persons whose usual source of care provider is a person in a facility.
    • Similarly, the universes for USCPRASN, USCPRBLK, USCPRHISP, USCPRNATAM, USCPROPI, USCPROTHRAC, USCPRWHT, and USCPRSEX have been updated to state that all persons with a usual source of care provider who is a person or a person in a facility are in universe. Previously, only persons with a usual source of care provider who is a person were listed as in universe.

February 19, 2020

    New data.

    • IPUMS MEPS now includes data from the Diabetes Care Survey (DCS) from 2000 to the present. For a full list of the variables included in the supplement, please see DIACONF. COVERTYPE and COVTYPEAGE, which measure health insurance coverage type and health insurance coverage type by age, respectively, have also been added to IPUMS MEPS.

    Corrected data.

    Modified data.

    • The codes for the variable RULETRRD in 2016 were updated. Although RULETRRD still uniquely identified a Reporting Unit within a Dwelling Unit in 2016, we updated values of "X" to "B" for greater consistency with other years of data.

December 3, 2019

    Modified data.

    • We altered the procedures we used to construct MEPSID and MEPSIDRD for panel 9 members in the 2005 sample only. Their MEPSID and MEPSIDRD values are now "0" + DUID + PID + PANEL. Previously, their MEPSID and MEPSIDRD values were DUID + PID + "0" + PANEL.

October 3, 2019

    Corrected data.

    • We corrected an error in the values of PSUANN in 2002. Previously, PSUANN contained the values of STRATANN by mistake.

September 25, 2019

    New data.

    • IPUMS MEPS now includes 2017 FYC data; over 900 variables (both annual person-level variables and round-level variables) are now available through the extract system.

    Modified data.

    • The codes for the variable ADPRXY were updated. The new codes accommodate the more general coding structure used by AHRQ in 2017. Please see the codes page for more information.
    • The width for the variable HIEU was changed from 12 to 7 wide after the elimination of right-side Bs in 1999 and 2000 only. HIEU previously did and continues to function as a unique identifier.

    Corrected data.

    • We fixed an error that incorrectly collapsed family identifiers in the 1996 and 2012 data for the variable FAMIDRD; family identifier values of "C" and "D" were combined into a single category ("C") and, in 2012 only, values of "E" in the original data were incorrectly coded as "D". This has been corrected. Additionally, in 1997, 1998, and 2016 we updated cases of FAMIDRD values of "X" to "B" for greater consistency with other years of data; the previous assignment of these years was still appropriate for unique identification, but the code has been modified.
    • We fixed an error that incorrectly reported responding unit identifiers in the 2012 data for the variable RULETRRD: responding unit identifier values that were "A" in previous versions of the data are now "C"; cases that previously had values of "I" are now "A"; cases that previously had values of "C," "D," and "E" were incorrect by one letter and are now assigned values of "D," "E," and "F" respectively. No values were collapsed, so all responding units were still uniquely identifiable using these codes. Additionally, in 1997 and 1998 we updated cases of RULETRRD values of "X" to "B" for greater consistency with other years of data; the previous assignment of these years was still appropriate for unique identification, but the code has been modified.

May 20, 2019

    Modified data.

    • The codes and value labels for a number of variables using a likert scale of "None of the time" to "Most of the time" in the self-administered questionnaire were modified for consistency throughout the entire questionnaire section. "None of the time" is consistently assigned a value of "0", while "All of the time" is now consistently assigned a value of "4". This change affets the following variables: ADDOWN, ADPALS, ADPWLM, ADMALS, ADMWLM, ADCAPE, ADNRGY, and ADSOCA. The three point likert scale for the variables ADDAYA and ADCLIM was also modified to reflect this general pattern, though their scale ranges from "Not limited" to "Limited a lot". Out of universe cases for these variables are assigned to the code of "6."

    Corrected data.

    • The value labels for the variable ADPAIN were incorrect. ADPAIN reports how often pain interefered with normal work in the past 4 weeks and reports values on a likert scale of "None of the time" to "Most of the time"; the value labels of the likert scale were incorrectly inverted. The codes and labels have been updated to reflect the correct direction of the likert scale, and to be consistent with other self-administered questionnaire items using this likert scale.

April 20, 2019

    Corrected data.

    • The top-codes for WAGEHRRD and WAGENEWRD have been corrected. WAGEHRRD previously assigned all top-coded values to the 1997 top-code value of $43.75, and WAGENEWRD previously assigned all top-coded values to the 2015 top-code value of $80.00. Data now report the appropriate year-specific top code for both variables.

November 27, 2018

    Corrected data.

    • We have inserted a leading zero in the values of MEPSID and MEPSIDRD in the 1996-2004 samples to replace a leading blank that was preventing users from merging data using formatted data extracts to ASCII data extracts read in using a stats package syntax file. The merging issue previously observed has been resolved.

    New data.

    • The variable PERNUM is now included on round-level records to support hierarchical IPUMS extracts in R.

October 4, 2018

    Modified data.

    • Corrected duplicate cases for 1996 sample.

August 31, 2018

    New data.

    • IPUMS MEPS now includes 2016 FYC data; over 1,300 variables (both annual person-level variables and round-level variables) are now available through the extract system.

    Modified data.

    • To account for the release of pooled variance estimation variables (STRATAPLD and PSUPLD); the variable names for STRATAP and PSUP were updated to STRATANN and PSUANN to more clearly note that they are designed to produce annual estimates.
    • A missing data code for HPTOTNIGHT was updated for consistency with other IPUMS missing data codes; cases previously coded as -9 are now classified as "998: Not ascertained".
    • Missing and top-code values for WAGENEWRD were updated for consistency with other IPUMS missing data codes (they previously matched AHRQ missing and top-coded values and were not labeled in data extracts). Cases previously coded as -1.00 are now classified as "0.00: Not in Universe"; cases previously coded as -10.00 are now classified as the respective top-code for each year of data; cases previously coded as -13.00 are now classified as "99.91: Initial wage imputed"; and cases previously coded as -8.00 are now classified as "99.98: Not ascertained".
    • The variable WAGEHRRD was expanded by one column and its missing data and top-code values were updated for consistency with other IPUMS missing data codes (they previously matched AHRQ missing data codes and were not-labeled in extracts). Cases previously coded as -1.00 or -3.00 are now classified as "0.00: Not in Universe"; cases previously coded as -2.00 are now classified "9999.94: Determined in previous round"; cases coded as -10.00 are now classified as the respective top-code for each year; cases previously coded as -7.00 are now classified as "9999.97: Refused"; cases previously coded as -8.00 are now classified as "9999.99: Don't know"; and cases previously coded as -9.00 are now classified as "9999.98: Not ascertained".
    • The missing data codes for HOURWRKRD in 2001, 2004, and 2007-2009 reflected AHRQ-assigned values, and were not consistent with other IPUMS missing data codes. Cases previously coded as -1 or -3 are now classified as "996: Not in Universe"; cases previously coded as -2 are now classified "994: Determined in previous round"; cases previously coded as -7 are now classified as "997: Refused"; cases previously coded as -8 are now classified as "999: Don't know", and cases previously coded as -8 are now classified as "998: Not ascertained".
    • The code for "determined in previous round" was changed from 96 to 94 for the round-level variable WAGEHRSRCRD for greater comparability across variables. The variable is otherwise unchanged.

August 13, 2018

    New data.

    • Newly released 2015 MEPS data and 300+ additional variables, including round-level, are now integrated into the IPUMS MEPS database.

    Modified data.

    • EDUCATN has been renamed EDUCYR to match the original MEPS variable name and additional samples were added.

June 21, 2018

    Updated data.

    • IPUMS MEPS has updated USBORN and YRSINUS (now named YRSINUSG in 2013 to include both panels 17 and 18. Previously, panel 17 members had been classified as "inapplicable" because of a change in how MEPS offers these variables.

    Corrected data.

    • IPUMS MEPS has corrected an error affecting CPSFAMID. In previous versions of the data, CPSFAMID used incorrect original data for 2008 and 2012, causing the data to be missing in 2012 and incorrect in 2008.

February 5, 2018

    Corrected data.

    • IPUMS MEPS has corrected an error affecting PID and MEPSID. In previous versions of the data, all values of PID and the PID portion of MEPSID (DUID + PID + PANEL) were zero.

January 30, 2018

    New data.

    • Beta version of IPUMS MEPS released.