Medical Services Expenditures by Disease: Diseases of the Circulatory System, MEPS Account Basis (DICISYEXPMEPS) | FRED | St. Louis Fed
- Rows
- 22
- Columns
- 2
- Data range
- 2000-01-01 to 2021-01-01
Graph and download economic data for Medical Services Expenditures by Disease: Diseases of the Circulatory System, MEPS Account Basis (DICISYEXPMEPS) from 2000 to 2021 about disease, healthcare, medical, health, expenditures, services, and USA.
AI analysis
Data Analysis: Medical Services Expenditures by Disease — Circulatory System (DICISYEXPMEPS)
1. Dataset Overview and Value for Correlation Studies
This dataset, sourced from the Federal Reserve Bank of St. Louis (FRED) via fredgraph.csv?id=DICISYEXPMEPS, tracks Medical Services Expenditures for Diseases of the Circulatory System using the Medical Expenditure Panel Survey (MEPS) account basis. The data spans 2000 to 2021 (22 annual observations), a 22-year window that captures multiple economic cycles, healthcare policy shifts (e.g., ACA implementation in 2010–2014), and demographic trends in cardiovascular disease treatment costs. As a FRED-hosted series, it benefits from the St. Louis Fed's standardized curation, making it a reliable, publicly citable source for economic and health-policy research. Its narrow focus (a single disease category, in likely nominal USD terms per MEPS conventions) makes it most valuable as a dependent or comparison variable in correlation studies examining healthcare spending drivers rather than a standalone analytical dataset.
2. Data Quality Observations
The dataset is clean by conventional metrics: zero null cells across both columns, and both observationdate and DICISYEXPMEPS show 22 distinct values matching the 22 total rows — indicating no duplicate timestamps or repeated expenditure values. There is no disproportionate null concentration in either column. One caveat: the duplicate-row check is flagged as "pending Phase D recomputation," so full duplicate-row confirmation is not yet finalized, though the distinct-count parity (22 distinct dates = 22 distinct values = 22 rows) strongly suggests no duplication exists. The sample data rows are not yet populated (Phase D/Parquet pending), which limits row-level spot-checking at this stage.
3. Key Column Distributions
The DICISYEXPMEPS column (expenditure values) ranges from 161 to 295.18, with a mean of 235.22 and median of 249.82 — the mean sitting notably below the median, consistent with the negative skew (-0.534), indicating a longer left tail (likely driven by lower expenditure values in earlier years, e.g., 2000s, pulling the average down relative to more recent, clustered higher values). The interquartile range (Q1=196.75, Q3=266.37) spans about 70 units, meaning the middle 50% of observations are fairly compressed compared to the full 134-unit range — suggestive of a steady upward trend over time with a plateau or deceleration in later years rather than uniform annual growth. With only 22 points, formal outlier detection (e.g., IQR fencing) isn't very robust, but the spread between the median and max (295.18) suggests recent years cluster near the top of the distribution — a classic "level shift up over time" pattern typical of nominal healthcare cost series.
4. Recommended Join Key
observationdate is the clear and only viable join key — it's annual (January 1 of each year), non-null, and uniquely identifies each row. For cross-dataset correlation, this should be matched against other annual time series (not monthly/quarterly data without aggregation first). Analysts should confirm date alignment (calendar year vs. fiscal year) and consider resampling higher-frequency datasets to annual means/sums before joining.
5. Suggested Complementary Datasets
Strong pairing candidates include: (a) other MEPS disease-expenditure series (e.g., neoplasms, diabetes, musculoskeletal) to compare relative disease-cost growth rates; (b) CPI for Medical Care Services (FRED: CPIMEDSL) to deflate nominal expenditures into real terms; (c) National Health Expenditure (NHE) data from CMS for macro-level context; (d) cardiovascular disease prevalence/mortality rates (CDC WONDER or NHANES) to test whether spending correlates with disease incidence versus treatment cost inflation; and (e) demographic/aging population data (e.g., % population 65+, FRED series) since circulatory disease burden is strongly age-correlated. Given the short annual series, correlation analyses should account for autocorrelation/trend effects (e.g., via differencing) rather than naive Pearson correlation on levels.
Columns
- observation_date (date)
- DICISYEXPMEPS (decimal)