How to check why your therapy practice reports disagree

Write down what each number counts

Start with what the report counts. A total called “sessions” might count appointments, each person attending a group, completed notes or items billed. A total called “revenue” might mean amounts billed or money received. The same label does not guarantee the same calculation.

For each important number, write down what it counts, which dates it uses and which people or appointments it includes. For a percentage, write the two numbers being compared. For money, also check refunds, reduced charges and any fees taken out. This short note lets the next person check the number the same way.

For example: “Weekly completed appointment events, assigned to the clinician who delivered the appointment, by service date, excluding cancelled and no-show events, with pending documentation tracked separately.” That is specific enough to test. “Weekly productivity” is not.

Check these differences in order

Check the items below before searching every spreadsheet formula. If two reports use different dates or include different clinicians, the difference may have a simple explanation.

1. Match the people

   Use the same clinicians and locations. Check whether staff access hides any records.

2. Check what a row counts

   One group appointment can produce several attendance or billing rows.

3. Use the same date type

   Session, claim and payment dates can place the same work in different months.

4. Check the status

   “Complete” may refer to care, documentation or payment.

5. Check clinician assignment

   A report may use the treating clinician or the client’s current clinician.

6. Compare the update times

   Corrections and late payments can change an earlier period.

Example: one appointment can create several rows

SimplePractice documents one report row per client in a group appointment. Sessions Health distinguishes completed appointments awaiting notes from completed notes. Those are concrete reasons to investigate units and statuses before assuming missing work.

Here is an example with invented totals. Report A shows 120 appointments. Report B has 132 rows. Four group appointments each had four clients. Those groups make 16 rows in B but only four appointments in A. The extra 12 rows explain the difference: 16 − 4 = 12, and 132 − 120 = 12.

Do not delete twelve rows to make the total match. Keep the information about who attended, and create a separate count of distinct appointments. Use attendances when you need to know how many people were seen. Use appointment events or time when you need to understand the clinician’s schedule.

Once the totals match for the agreed unit, test a second period and a clinician with a different service mix. That checks whether you fixed a rule rather than fitted a one-time explanation.

Example: this month’s payments can cover older sessions

Suppose September services eventually produce $50,000 of collections: $35,000 received in September and $15,000 in October. September also contains $12,000 received for August services. September cash is therefore $47,000, while eventual collections for September services are $50,000. The $3,000 difference is explained by timing: $15,000 still to arrive minus $12,000 from earlier care.

Both totals can be right. The money received in September helps you plan September’s bills. Payments for September’s sessions help you understand what that care eventually earned. Dividing September’s payments by September’s sessions mixes the two if some money belongs to other months.

Also check refunds, changes to charges and money that has not yet been matched to a particular bill. Keep these differences visible. Putting them into an unexplained “other” amount makes the next check harder.

Decide how to handle later corrections

Decide whether a saved report should stay as it was or update when records change. A saved copy shows what you knew on that date. A refreshed report may include later payments or corrections. Both are useful, but do not compare them as though they contain the same information.

Keep a dated copy of the owner review and distinguish it from the current source report. When a material historical number changes, record the cause and whether the comparison was restated. A payment received later is different from a corrected appointment status, and each affects a different management question.

When a percentage changes, check both numbers behind it. For example, 8 of 10 and 8 of 12 have the same first number but different percentages. Showing the counts helps people understand the change, especially in a small practice or a short review period.

Fix the cause of repeated differences

Keep a short list of differences: the report, dates, cause, correction and the person checking the fix. If the same issue could happen again, add a simple safeguard. That might be agreed report settings, a clear clinician assignment or a check of which services a payment covers.

There are two kinds of fix. Sometimes the original record is wrong and needs a correction. Sometimes the record is right but you need a different calculation, such as counting appointments instead of attendances. Make clear which you are doing. Adding a number to force a match does not explain either one.

Use invented example records when testing outside approved systems. You can test the method without putting patient names into a public spreadsheet or website form.

Use these checks when setting up Cortexa

Cortexa Analytics helps owners review sessions, revenue, clients, clinician activity and filled appointment time. During setup, agree what each important number means and check a complete period against the original records. That gives the practice a reliable starting point for later comparisons.

If Full Service Partner also supports the practice, the finding can point to a task Cortexa handles, such as intake information, billing follow-up or payroll. The report shows the difference; the person responsible fixes its cause. That helps avoid the owner correcting the same issue every week.