Professional Services Review Scheme Directions as to Sampling (No. 2) (Cth)

Health Insurance Act 1973

Professional Services Review Scheme –

Directions as to Sampling (No. 2)

I, CARMEN MARY LAWRENCE, Minister for Human Services and Health, acting under section 106K of the Health Insurance Act 1973, issue the following Directions, being directions that are in accordance with advice received from the Australian Bureau of Statistics, for the purposes of  sections 106H and 106J of that Act.

Dated 31 January 1995

CARMEN LAWRENCE

Minister for Human Services and Health

_____________

PART 1 – PRELIMARY

1.1   This Direction may be cited as Professional Services Review Scheme – Directions as to Sampling (No. 2).

1.2 This Direction sets out procedures in regard to samples of services for the purposes of Subdivision C (Findings based on statistical sampling) of Division 4 of Part VAA of the Health Insurance Act 1973 (the Act).

1.3   Terms used in this Direction that are defined in Part VAA of the Act have the same meanings in this Direction as in that Part.

1.4   Part 2 of this Direction sets out the procedure for producing and using a sample for the purposes of subsection 106H(1) of the Act.  Part 3 sets out the procedure for increasing the size of a sample if the practitioner under review so requests under subsection 106J(2) of the Act. 

Professional Services Review Scheme – Directions as to Sampling

1.5   A sample produced for any purpose referred to in these Directions must be randomly selected in accordance with proper statistical sampling practice unless the contrary is specifically stated.

PART 2 – SAMPLING

2.1   If a referral to the Director relates to services of a class specified in the referral (“the specified class”) that may be the subject of findings under Division 4 of Part VAA of the Act, the Commission may supply a sample  (“preliminary random sample “) drawn from the services included in the referral.  The preliminary random sample must contain twice the number of services worked out by reading from Table 1 the number at the intersections of:

(a)the row containing in its leftmost column (the column headed “class size”) the number of services of the specified class included in the referral; and

(b)the column headed “50%”.

2.2   If the Committee proposes to make a finding based on statistical sampling, the Committee must examine a sample of 30 services (“exploratory sample”) from the preliminary random sample and determine whether each of those services constitutes inappropriate practice or not.  The Committee must then calculate the percentage of services in the exploratory sample that constitute inappropriate practice and round down the result to the next lower multiple of 5%.

2.3   If the percentage worked out in the way set out in clause 2.2 (“rounded percentage”) is less than 20%, no further statistical inferences are to be made concerning the referral.

2.4   If the rounded percentage is 20% or greater, the Committee must examine further cases from the preliminary random sample until it has examined the number of cases found in Table 1 at the intersection of:

(a)the row containing in its leftmost column (the column headed “class size”0) the number of services of the specified class included in the referral; and

(b)the column headed by the rounded percentage.

This sample (“final random sample”) may include the services that were included in the exploratory sample.

2.5    The Committee must determine whether each service included in the final random sample constitutes inappropriate practice, and work out the percentage of services in that sample that constitutes inappropriate practice.

Professional Services Review Scheme – Directions as to Sampling

2.6    The percentage worked out in the way described in clause 2.5 must then be reduced by ten percentage points.  The resulting percentage must be taken to be the percentage of services of the same kind as the services included in the referral that constitute inappropriate practice.

[NOTE:  Call the percentage worked out as set out in clause 2.5 “x”. To a confidence level of 95%, the percentage of services in the sample that constitutes inappropriate practice is within the interval (x ± 10)%.  For the purposes of the Committee’s findings, the lower limit of that interval is to be taken.]

PART 3 – INCREASING THE SAMPLE SIZE

3.1    If a Committee proposes to make findings based o the final random sample of services, and the person under review requests that the sample size be increased, the Committee must select from the preliminary random sample a further sample containing the same number of cases as the number of cases that was included in the final random sample used in the committee’s first consideration of the referral.  The latter sample is to be combined with the final random sample to make up the ‘double final random sample”.

3.2    The Committee must determine whether each service included in the doubled final random sample constitutes inappropriate practice, and work out the percentage of services in that sample that constitutes inappropriate practice.

3.3    The percentage worked out in the way described in clause 3.2 must then be reduced by half the number of percentage points found in Table 2 at the intersection of:

(a)    the row containing in its leftmost column (the column headed “class size”) the number of services of the specified class included in the referral; and

(b)    the column headed by the percentage calculated in clause 3.2 rounded down to the next lower multiple of 5%.

The resulting percentage must be taken to be the percentage of the services included in the referral that constitute inappropriate practice.

[NOTES:

1.      Call the percentage worked out as set out in clause 3.2 “x”, and the percentage found from Table 2 “y”.  To a confidence level of 955, the percentage of services in the sample that constitutes inappropriate practice is within the interval x ± .  For the purposes of the Committee’s findings, the lower limit of that interval is to be taken.

Professional Services Review Scheme – Directions as to Sampling

2.      An example of the use of the process follows.  This example is intended to be an illustration only.

Assume the Commission includes in its referral 500 surgical excisions under the same item number.  It will also include 168 of those excisions randomly selected – 84 being the size of the random sample necessary to predict with 95% confidence that a result derived from the sample will be within ±10% of the actual position in the population at a hypothesised incidence of inappropriate practice of 50% (the incidence requiring the largest sample).

The Committee will consider the first 30 such services (the exploratory sample) and make a finding as to what number, if any, of those services were inappropriate.

If the Committee decides that less than 20% (6 excisions) were inappropriate, no conclusion can be drawn as the result will not be statistically valid.

If the committee decides that 6 or more of the services were inappropriate, say 8 or 26.7%, the 26.7% is rounded down to the nearest 5%, that is 25%.  From Table 1, a sample size of 66 is determined.

The Committee then considers a further 36 excisions from the random sample (the number needed to make up the final random sample) and if it finds that 18 or 27% of the 66 services it has now considered were inappropriate, it can find that 17% (27% minus the possible 10% error) or 85 of the excisions were inappropriate.

At this point the practitioner has effectively three options:

• accept a finding that 17% of the excisions were inappropriate; or

• require the Committee to consider a further 66 randomly selected excisions, thus doubling the sample size and reducing the confidence interval, so that if a similar pattern were found across the whole 132 excisions the confidence interval would drop to 13% giving an accuracy of ± 6.5% - the new confidence interval derived from Table 2.  The finding then would be that 20.5% or 102.5 of the excisions were inappropriate practice.  There is also the possibility that the larger sample might demonstrate that the correction should have been the other way and result in a higher percentage of inappropriate excisions, theoretically as high as 30.5% (37% minus 6.5%); or

• take the Committee through all 500 excisions. The effect is that, if the sample is accurate, 27% or 135 inappropriate excisions would be the expected result.]

PART 4 – REVOCATION

4.1 The Directions entitled “Professional Services Review Scheme – Directions as to Sampling”¹, issued under section 106K of the Health Insurance Act 1973 on 9 May 1994, are revoked.

[NOTE:

1. The Directions were notified in the Commonwealth of Australia Gazette on 25 May 1994.]

Professional Services Review Scheme – Directions as to Sampling

TABLE 1

SAMPLE SIZES FOR A 95% CONFIDENCE INTERVAL

OF ± 10%

 

Class size	Percentages of services that represent inappropriate practice
20%	25%	30%	35 %	40%	45%	50%	55%	60%	65%	70%	75%	80%	85%	90%
50	29	30	32	33	33	34	34	34	33	33	32	30	29	26	22
100	40	43	46	48	49	50	51	50	49	48	46	43	40	34	27
150	46	51	54	57	59	60	61	60	59	57	54	51	46	39	30
200	49	55	60	63	66	67	67	67	66	63	60	55	49	41	31
250	52	58	64	67	70	72	72	72	70	67	64	58	52	43	32
300	54	61	66	71	73	75	76	75	73	71	66	61	54	44	33
350	55	63	69	73	76	78	79	78	76	73	69	63	55	45	34
400	56	64	70	75	78	80	81	80	78	75	70	64	56	46	34
450	57	65	72	77	80	82	83	82	80	77	72	65	57	47	34
500	58	66	73	78	81	83	84	83	81	78	73	66	58	47	35
550	58	67	74	79	83	85	85	85	83	79	74	67	58	48	35
600	59	68	75	80	84	86	87	86	84	80	75	68	59	48	35
650	59	68	75	81	85	87	88	87	85	81	75	68	59	48	35
700	60	69	76	81	85	88	88	88	85	81	76	69	60	48	35
750	60	69	76	82	86	88	89	88	86	82	76	69	60	49	35
800	60	69	77	83	87	89	90	89	87	83	77	69	60	49	35
60	70	77	83	87	90	90	90	87	83	77	70	60	49	35
900	61	70	78	84	88	90	91	90	88	84	78	70	61	49	36
950	61	70	78	84	88	91	91	91	88	84	78	70	61	49	36
1000	61	71	78	84	89	91	92	91	89	84	78	71	61	49	36
1500	62	72	80	87	91	94	95	94	91	87	80	72	62	50	36
2000	63	73	32	88	93	95	96	95	93	88	82	73	63	51	36
2500	63	74	82	89	93	96	97	96	93	89	82	74	63	51	36
3000	64	74	83	89	94	97	98	97	94	89	83	74	64	51	37
3500	64	74	83	90	94	97	98	97	94	90	83	74	64	51	37
4000	64	75	83	90	95	98	99	98	95	90	83	75	64	51	37
4500	64	75	83	90	95	98	99	98	95	90	83	75	64	51	37
5000	64	75	84	90	95	98	99	98	95	90	84	75	64	51	37

Professional Services Review Scheme – Directions as to Sampling

TABLE 2

95% CONFIDENCE INTERVAL WIDTHS FOR SAMPLES TWICE THE

SIZE OF THOSE IN TABLE I

Class size	Percentages of services that represent inappropriate practice
20%	25%	30%	35%	40%	45%	50%	55%	60%	65%	70%	75%	80%	85%	90%
50	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%
100	8%	7%	5%	4%	2%	0%	0%	0%	2%	4%	5%	7%	8%	10%	11%
150	11%	10%	9%	9%	8%	8%	8%	8%	8%	9%	9%	10%	11%	11%	12%
200	12%	11%	11%	10%	10%	10%	10%	10%	10%	10%	11%	11%	12%	12%	13%
250	12%	12%	11%	11%	11%	11%	11%	11%	11%	11%	11%	12%	12%	13%	13%
300	12%	12%	12%	12%	12%	11%	11%	11%	12%	12%	12%	12%	12%	13%	13%
350	13%	12%	12%	12%	12%	12%	12%	12%	12%	12%	12%	12%	13%	13%	13%
400	13%	13%	13%	12%	12%	12%	12%	12%	12%	12%	13%	13%	13%	13%	13%
450	13%	13%	13%	13%	12%	12%	12%	12%	12%	13%	13%	13%	13%	13%	13%
500	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%
550	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%
600	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%
650	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%
700	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%
750	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%
800	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%
850	14%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%	14%
900	14%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%	14%
950	14%	14%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%	14%	14%
1000	14%	14%	13%	13%	13%	13%	13%	13%	13%	13%	13%	14%	14%	14%	14%
1500	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
2000	14%	14%	14%	14%	14%	14%	1-1%	14%	14%	14%	14%	14%	14%	14%	14%
2500	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
3000	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
3500	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
4000	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
4500	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%	14%
5000	14%	14 %	14%	14%	14%	14%	14%	1.4%	14%	14%	14%	14%	14%	14%	14%

________________________________________________________________