| Title: | Tools for Obtaining and Cleaning Medicare Public Use Files |
|---|---|
| Description: | Publicly available data from Medicare frequently requires extensive initial effort to extract desired variables and merge them; this package formalizes the techniques I've found work best. More information on the Medicare program, as well as guidance for the publicly available data this package targets, can be found on CMS's website covering publicly available data. See <https://www.cms.gov/Research-Statistics-Data-and-Systems/Research-Statistics-Data-and-Systems.html>. |
| Authors: | Robert Gambrel [aut, cre] |
| Maintainer: | Robert Gambrel <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.2.1 |
| Built: | 2024-11-14 03:13:19 UTC |
| Source: | https://github.com/robertgambrel/medicare |
Provide names for Cost Report "Alpha Table" data
cr_alpha_names()cr_alpha_names()
A list of names for the cost report Alpha Table
# get the list cr_nmrc_names()# get the list cr_nmrc_names()
This function takes a 5-column alpha-numeric dataset or numeric dataset from the Medicare cost reports, which are stored in a long format, and subsets them based on the worksheet number, line number, and column number provided. If desired, it will rename the resulting variable to whatever the user chooses.
cr_extract(dataset, worksheet, row, column, newname = "newvar")cr_extract(dataset, worksheet, row, column, newname = "newvar")
dataset |
The name of a cost report alpha or numeric dataset |
worksheet |
The name of the workheet, converted to 7-character format |
row |
The row number of the data, as it appears in the Medicare workbook or documentation (i.e. at least 3 digits. Row 5 must be entered as 500, row 5.1 as 501, etc.) |
column |
The column number of the data, as it appears in the Medicare workbook (same general rule as for rows) |
newname |
The name given to the variable that appears as a result of this extraction |
It does not automatically adjust for the same variable having different rows / columns in Medicare data formatted for the 1996 vs 2010 form. The user may have to use this function twice, once on each source of data, to extract one variable over time.
It does automatically recode rows and columns into all possible permutations (ie '500', '0500', '00500', 500) when subsetting, since different cost reports use different schema.
This function *does not* throw an error if the parameters yield an empty dataset at any point. It only gives warnings. This is because oftentimes the parameters are valid but the data is missing in the source material, due to CMS scrubbing of what data gets published.
A 2-column dataset: one with the cost report rpt_rec_number,
used to merge data, and a column of the data requested, which is renamed if
desired.
alpha_data <- hospiceALPHA hospice_name <- cr_extract(alpha_data, "S100000", 100, 100, "name")alpha_data <- hospiceALPHA hospice_name <- cr_extract(alpha_data, "S100000", 100, 100, "name")
Provide names for Cost Report "Numeric Table" data
cr_nmrc_names()cr_nmrc_names()
A list of names for the cost report Numeric Table
# get the list cr_nmrc_names()# get the list cr_nmrc_names()
Provide names for Cost Report "Report Table" data
cr_rpt_names()cr_rpt_names()
A list of names for the cost report Report Table
# get the list cr_rpt_names()# get the list cr_rpt_names()
A dataset containing the alpha data for the first 500 hospices in the Hospice
2014 cost reports. This is raw data, similar to what you'd get on your own
with read.csv("hospc_2014_ALPHA.csv", stringsAsFactors = FALSE).
hospiceALPHAhospiceALPHA
A data frame with 61820 rows and 5 variables:
V1The rpt_rec_num, used to link a hospices dataset across the 3 yearly files.
V2The wksht_cd, indicating which worksheet the variable comes from.
V3The line_num, indicating the line on the worksheet where the variable is found.
V4The clmn_num, indicating the column on the worksheet where the variable is found.
V5The itm_alphanmrc_itm_txt, indicating the variable's value.
A dataset containing the numeric data for the first 500 hospices in the Hospice
2014 cost reports. This is raw data, similar to what you'd get on your own
with read.csv("hospc_2014_NMRC.csv", stringsAsFactors = FALSE).
hospiceNMRChospiceNMRC
A data frame with 200,202 rows and 5 variables:
V1The rpt_rec_num, used to link a hospices dataset across the 3 yearly files.
V2The wksht_cd, indicating which worksheet the variable comes from.
V3The line_num, indicating the line on the worksheet where the variable is found.
V4The clmn_num, indicating the column on the worksheet where the variable is found.
V5The itm_val_num, indicating the variable's value.
A dataset containing the report data for the first 500 hospices in the Hospice
2014 cost reports. This is raw data, similar to what you'd get on your own
with read.csv("hospc_2014_RPT.csv", stringsAsFactors = FALSE).
hospiceRPThospiceRPT
A data frame with 500 rows and 5 variables:
V1The rpt_rec_num, used to link a hospices dataset across the 3 yearly files.
V2The prvdr_ctrl_type_cd, indicating the ownership structure of the facility.
V3The prvdr_num, a 6-character unique ID used to link the facility's information across time and with other Medicare data.
V4The npi, a unique provider number assigned under HIPAA, which can also be used to link to other data sources.
V5The rpt_stus_cd, the status of the report (initial submission, audited and settled, settled w/o audit, reopened). Facility filings can be revised, so even older years' data might be updated if data is accessed multiple times.
V6The fy_bgn_dt, the start date for the fiscal year of filing. Most facilities submit cost reports soon after close of their fiscal year. They can also have multiple entries in a calendar year if they change their fiscal year start and end dates.
V7The fy_end_dt, fiscal year end date. Usually 365 days after the start, unless the facility is re-basing its fiscal year system.
V8The proc_dt, process date, when the report was processed by CMS.
V9The initl_rpt_sw, initial report indicator, not currently actively used.
V10The last_rpt_sw, last report indicator, not currently used.
V11The trnsmtl_num, the current transmittal number when the report was generated.
V12The fi_num, fiscal intermediary number, which denotes which fiscal intermediary processes the facility's filings.
V13The adr_vndr_cd, automated desk reviewer vendor code, indicating the vendor for the fiscal intermediary.
V14The fi_creat_dt, when the fiscal intermediary processed the submitted report.
V15The util_cd, indicating the level of medicare utilization by the facility.
V16The npr_dt, the date of notice of program reimbursement.
V17The spec_ind, a CMS internal special purposes code.
V18The fi_rcpt_dt, the date the cost report was received by the fiscal intermediary.
The medicare package contains useful functions for manipulating raw Medicare public use files. These sometimes come with SAS read-in code provided, but more frequently require the analyst to manually recode and rename variables based on thorough review of the data documentation. This package focuses mostly on Cost Reports and Provider of Service files, but more support will be added for other sources in the future.
For more information on Medicare the the data available, see:
Provide pre-extracted names for Provider of Service file, years 2000-2010
pos_names(year)pos_names(year)
year |
A year in the range 2000-2010 |
A list of names for the POS dataset in the year specified, in the order that the raw data lists them
This function returns the results of running pos_names_extract on
the layout and raw data files for the chosen year. For years 2000-2010, the
raw data had unhelpful, generic, sequentially numbered variable names.
These can be calculated fresh by calling pos_names_extract on the
imported dataframe and its corresponding layout .txt file. Alternately,
this function returns the names compiled by that function for years
2000-2010, saving the user a step.
# get the list pos_names(2005) pos_names(2010)# get the list pos_names(2005) pos_names(2010)
This function takes a Provider of Services Record Layout file (in .txt form) and parses it to extract the descriptive variable names instead of generic ones. For example, the 2006 file variable PROV0085, which is the name of the variable in the raw dataset downloaded from CMS, has a more descriptive name in the layout file: CATEGORY-SUBTYPE-IND.
pos_names_extract(layout_file, data_file)pos_names_extract(layout_file, data_file)
layout_file |
The file location of the layout file |
data_file |
The year's data file |
This uses regular expressions to find variable names. It works with years
2000-2010. Later years seem to have descriptive names already, though they
aren't necessarily identical across years (nor do they match the names
produced here). This code can be run to produce variable names fresh, but
pre-computed variable names can also be accessed by names_pos_20XX()
functions also in this package.
A vector of names, ordered to match the corresponding year's data file
## Not run: pos_names_extract("pos_2006_layout.txt", pos_2006_data) ## End(Not run)## Not run: pos_names_extract("pos_2006_layout.txt", pos_2006_data) ## End(Not run)
A sample of Provider of Services data for select hospices in 2010. This is raw data, similar to what you'd get on your own when reading in a csv.
pos2010pos2010
A data frame with 402 rows and 530 variables
https://www.nber.org/data/provider-of-services.html
CMS publishes yearly final rules that detail annual price increases across various sectors of healthcare spending. In order to analyze spending increases due to utilization changes, it is frequently useful to "deflate" spending based on a reference period, so that observed changes are not due to inflation.
price_deflate(current_value, sector, current_year, reference_year = 2007)price_deflate(current_value, sector, current_year, reference_year = 2007)
current_value |
The current value that is being deflated to reference-period-equivalent dollars |
sector |
What sector is being adjusted. Currently supports: ip, op, phys, snf, hh, hospice, part_b_drugs, part_d_drugs, dme, and other |
current_year |
The current year (2007 - 2014) |
reference_year |
The base period to standardize to (2007 - 2014). |
Most sectors come from annual Federal Register Final Rules. Example: Physician Fee Change Rules.
Exceptions are other, which uses the general CPI deflator, and
part_d_drugs, which use the CPI-Pharmaceutical deflator.'
A float value, current_value / (current year index / reference year index)
# convert $100 in current inpatient spending to year 2007 dollars price_deflate(100, "ip", 2014, 2007)# convert $100 in current inpatient spending to year 2007 dollars price_deflate(100, "ip", 2014, 2007)
This function is not standalone - it is called from the cr_extract function.
subset_column(dataset, column)subset_column(dataset, column)
dataset |
The name of a cost report alpha or numeric dataset |
column |
The column of the workheet |
A subset of the provided dataset, subset to only having columns of the correct value
This function is not standalone - it is called from the cr_extract function.
subset_row(dataset, row)subset_row(dataset, row)
dataset |
The name of a cost report alpha or numeric dataset |
row |
The number of the row |
A subset of the provided dataset, subset to only having rows of the correct value
This function is not standalone - it is called from the cr_extract function.
subset_worksheet(dataset, worksheet)subset_worksheet(dataset, worksheet)
dataset |
The name of a cost report alpha or numeric dataset |
worksheet |
The name of the workheet, converted to 7-character format |
A subset of the provided dataset, subset to only having worksheets of the correct value