Antibiogram software vs spreadsheet for hospital pharmacy and lab teams
A worked example of the same culture export handled in Excel and in software, the five places spreadsheets go wrong, and the point where buying a tool is cheaper than the hours.
8 min read
A spreadsheet is fine for an antibiogram if one careful person builds it once a year from a clean export. Software earns its price when the duplicate isolate rule, the 30 isolate minimum, unit cuts or a second facility enter the picture, because that is where hand-built tables quietly go wrong and nobody can show a surveyor how a number was produced.
This is written for the pharmacy director, stewardship pharmacist or microbiology supervisor who inherited the Excel file and is deciding whether to keep it. There is a worked example, a comparison table and an honest note on when you should not buy anything.
A worked example with 612 isolates
Take a community hospital export of 612 E. coli results from one year. The raw rows include repeat cultures from the same patients. CLSI M39 asks for one isolate per patient per analysis period, the first one, so the real denominator is smaller. Suppose 140 rows are repeats. The antibiogram denominator is then 472 patients, not 612.
Now say 118 of the 612 rows are resistant to ciprofloxacin and 96 of the 472 first isolates are. The spreadsheet that counts every row reports 118 / 612 = 19.3 percent resistant, or 80.7 percent susceptible. The first isolate method reports 96 / 472 = 20.3 percent resistant, 79.7 percent susceptible. One point of difference looks small, but the repeat rows are usually from sicker patients who already failed therapy, so the direction of the error is not random and the gap grows in a unit like the ICU.
Where spreadsheets go wrong
- Duplicate isolates: a pivot table counts rows, not patients, so repeats inflate the denominator unless someone writes and maintains a de-duplication formula
- The 30 isolate rule: CLSI M39 advises against reporting percentages for species with fewer than 30 isolates, and Excel will happily print 100 percent susceptible on 6 isolates
- Intermediate results: a hand formula that counts only S and R drops the I category or counts it inconsistently
- Version drift: the file is copied for each committee, edited, and three versions disagree by the next meeting
- No audit trail: when a surveyor asks how a number was produced, the answer is the memory of whoever built it
Antibiogram software vs spreadsheet side by side
| Requirement | Spreadsheet | Antibiogram software |
|---|---|---|
| First isolate per patient | Custom formula you maintain | Built in, rule printed on the report |
| Fewer than 30 isolates | Manual check per cell | Flagged or suppressed automatically |
| Cuts by unit, source, quarter | New pivot per request | Filters on the same data |
| Second facility or a network | One file per site, merged by hand | Separate datasets, same method |
| Repeatability next year | Depends on one person | Same upload and mapping |
| Cost | Staff hours, no license | From $249 a month, hours drop |
How much does the spreadsheet really cost?
Count the hours from your own year. Cleaning the export, de-duplicating, building the pivots, formatting for the committee and answering two follow-up requests is a block of pharmacist or laboratory time. Multiply it by a loaded hourly rate and compare it with 12 months at $249, which is $2,988 a year for the Lab plan. If your cost is below that, keep the spreadsheet. If a stewardship pharmacist is doing it instead of reviewing cases, the trade is easy.
When you should keep Excel
Keep it if you report one species set once a year, the lab already de-duplicates the export, and a reviewer other than the builder checks the numbers. Follow the method in the CLSI M39 antibiogram guideline and print your rules on the page. Many small teams do this well. The line to cross is a second report cycle, a second site, or a surveyor asking for provenance.
A checklist before you decide
Answer five questions from your own last report. How many hours did it take from export to the printed table. Who checked the numbers other than the person who built them. Did the table apply the first isolate rule and say so. Were any percentages printed for species with fewer than 30 isolates. Could you rebuild last year's figure today from the same file. Two or more answers that you do not like is the signal that the spreadsheet has become a risk rather than a saving.
The second signal is scope. A request for cuts by unit, by source, or a second facility is where a single workbook multiplies into copies. Tools that keep one dataset and filter it avoid that. The cost side is covered in the antibiogram software cost breakdown, and a MRSA specific view is on the MRSA antibiogram page.
What switching looks like
You export a susceptibility line list from the LIS, upload it, map the columns once and review the output against last year's table. There is no interface or implementation project. The cumulative report is covered on the hospital antibiogram page, the M39 rules are listed on CLSI M39 antibiogram software, and plans are on the pricing page. If the question is also ESBL and carbapenem trends, see ESBL antibiogram software.
Related on this site: hospital antibiogram software, CLSI M39 antibiogram software, ESBL antibiogram software, antibiogram software cost route by route.