Huge difference between NPDB and CIHI abortion numbers

I continue to try and understand why the accuracy of abortion statistics in this country are so pathetic. I contacted CIHI to get more clarifications about their numbers.

As we already know CIHI's data sources are here:

1. Discharge Abstract Database, Canadian Institute for Health Information.

2. National Ambulatory Care Reporting System, Canadian Institute for Health Information.

3. National Physician Database NPDB (here), Canadian Institute for Health Information.

4. Ministère de la Santé et des Services sociaux du Québec.

5. Manitoba Health, Seniors and Active Living.

6. Induced abortion clinics in Newfoundland and Labrador, Saskatchewan and British Columbia.

When I compared the numbers in the National Physician Database (number 3 above) I learned that these numbers are very different from CIHI's quick stats (which is what everyone uses for yearly abortion numbers).

So I asked CIHI to explain the difference between the two sets of numbers, since the quick stats use a bunch of sources, and the NPDB uses only one source? 

Here is the response I received from CIHI:

"CIHI’s annual induced abortion tables report on abortion volumes in Canada. It includes data on hospital-based abortions submitted to national databases (e.g. DAD, NACRS), in addition to clinic data that is voluntarily provided to CIHI. These tables also include NPDB data as a source for Ontario to estimate the volume of non-hospital abortions.

The NPDB annual release includes NPDB data only, with the goal of national reporting on payments to physicians and services provided. The “utilization” data tables that report service volumes are based on fee-for-service data only. Services provided under alternative forms of reimbursement (e.g. salary and capitation) are not included.

In addition, different methodologies are used to identify abortion volumes between the two reports (Canadian Classification of Health Interventions, CCI, for hospitalization data, vs Fee for service billing codes in NPDB).

Given these differences, there will be expected differences in volumes."

In other words, the NPDB uses fee for services numbers only. And different (undefined) methodologies are used. I've asked what these methodologies are.

Shouldn't fee for service data be more accurate? Why isn't it? Why do they need to obtain the data from so many different places? Then have to use methodologies to get something resembling accurate stats? Why not just make the reporting for abortions performed mandatory and insist every province provide that info?

One can only imagine the gyrations CIHI goes through to come to numbers that they hope to, but never actually achieve, be accurate.

CIHI is still under-reporting abortion statistics

Abortion statistics are getting more confusing.

The Ministry of Health and Long-Term Care (MOHLTC) reported a total of 44,430 abortions for 2015/2016.

CIHI reported a total of 39,679 abortions for 2015.

That's a total difference of 4,751 abortions for one year.

To add to the confusion, MOHLTC reports using seven abortion codes (S752A, S785A, P054A, S770A, S783A, P001A, A920A), while CIHI reports using only two abortion codes (S752A and S785A):

"The enhanced methodology implemented in CIHI’s 2015 report for estimating the volume of induced abortions in Ontario incorporates NPDB (1) data selected using two fee codes for surgical abortions only. The report you shared includes seven fee codes and CIHI reports using two....the data sources for CIHI’s report are DAD, NACRS, and NPDB, while it appears that the Claims History Database was the source for the report you shared [from the MOHLTC]. (2)

If we only compare CIHI with MOHLTC for the same two codes that CIHI reports, we get CIHI at 39,679 abortions, and MOHLTC at 37,806 abortions. for a difference of 1,873 more than MOHLTC.

Neither CIHI nor MOHLTC could explain these differences.

Therefore CIHI reports 4,751 overall less than MOHLTC, but reports 1,873 more than MOHLTC if we only look at the same two codes.

Very confusing I know.

MOHLTC full fee code descriptions here.
___________________________________________________________________________

Additional Notes

Both CIHI and MOHLTC use these two codes:
S752A and S785A (surgical abortions)

The five codes not used by CIHI data, but are used in MOHLTC data are:
A920A (MEDICAL MANAGEMENT OF EARLY PREGNANCY - 5933)
P001A (MEDICAL MGMT FETAL DEMISE BETWEEN 14‐20 WKS GESTATION - 556)
P054A (fetal reductions - 115)
S770A CORPUS UTERI‐HYSTEROTOMY (16)
S783A CORPUS UTERI‐HYSTEROTOMY WITH TUBAL INTERRUPTION (4)

Those five additional codes used at MOHLTC add up to 6,624 abortions (115+16+4+556+5933). The Ministry also tells us that the S770A and S783A codes can be used for procedures other than abortions, but give me no indication as to which of these codes are not abortions. When I asked the Ministry to clarify medical abortions and identify which diagnostic codes are used for them, I was informed that:

"The OHIP claims system does require a physician to submit a fee code for payment however it is optional for a physician to submit a diagnostic code with the fee code and there is no rule that the diagnostic code needs to relate to the fee code."

Therefore it looks like the diagnostic codes are useless. So until the Ministry can tell me otherwise, I will assume that all of these additional procedures are also for abortions.

A920A and P001A codes are both used for medical abortions (556+5933=6,489 medical abortions), and CIHI isn't capturing these numbers. CIHI is also not capturing P054A for fetal reductions (115)

Note that CIHI doesn't report "medical" abortions or "fetal reduction" abortions.

That's 6,624 additional abortions being performed in Ontario for 2015/2016 (i.e. the five missing codes), than is officially being reported by CIHI.

(1) "The National Physician Database (NPDB) is a CIHI database that contains physician billings for publically funded insured medical services. Physicians do not report to it directly.  Physicians submit claims to the Provincial/Territorial medical care plans to be paid.  All provinces submit data from their claims systems to the National Physician Database." personal correspondence July 12, 217 with CIHI
(2) personal correspondence June 23, 217 with CIHI