Table 2 HTA report with data for each decision criterion of the EVIDEM framework ('Lite' highly synthesized version)*

Disease:

Type of intervention: Cervical cytology test

Liquid Based Cytology (LBC)

Intervention:

Indication: Screening for cervical cancer in women

Cervical cancer

Setting:

Administration: Liquid based cytology (LBC) requires physician to obtain

Discovery Health, South Africa

cells from the cervix and to place the head of the spatula/brush or rinse it into a vial with preservative fluid. The pathologist then extracts the cells into a microscopy slide and the cells are examined in the usual way

Comparator(s): Conventional pap smears (i.e., conventional cytology, CC) where the brush containing the sample is smeared onto the slide, thereby transferring the cells

Data included: Data available from public domain

MCDA matrix criteria

Highly synthesized information

Scoring of intervention

Disease impact

Not severe

Very severe

D1

Disease severity

Disease symptomatic only when spreading from cervix: vaginal bleeding, post-coital spotting, vaginal discharge, pelvic or low back pain.

0

1

2

3

Late stages: severe anemia, weight loss and uncontrolled release of urine and feces through the vagina

Survival: < 18 months in 50% of untreated patients

D2

Size of population

Incidence in South Africa: 9/100,000 white women; 40/100,000 black women

Very rare disease

Very common disease

0

1

2

3

Intervention

I1

Clinical guidelines

LBC is not included in any guidelines.South African guidelines recommend at least 3 pap smears test per lifetime

Not recommended

Strong recommendation

0

1

2

3

I2

Comparative interventions limitations

Sensitivity: 74% - relatively low; Specificity: 87%;Unsatisfactory sample: 3%Low screening uptake: 2.8% black women; 18.8% white women

No or very minor limitations

Major limitations

0

1

2

3

I3

Improvement of efficacy/effectiveness

Meta-analyses of RCTs (n = 28,736 vs 39,377): Sensitivity: 80% - difference with CC = 6.4% (95% CI: -6.5 to 18.8%)Specificity: 82% - difference with CC = - 4.0% (95% CI: -19.9 to 10.6%) Unsatisfactory samples: ThinPrep = 2.2% and Sure Path = 0.82%; difference with CC: ThinPrep = -0.8% and Sure Path = -2.5%

Lower than comparators

Major improvement

0

1

2

3

I4

Improvement of safety & tolerability

Safety and tolerability do not differ between LBC and conventional cytology

Lower than comparators

Major improvement

0

1

2

3

I5

Improvement of patient reported outcomes

Patient reported outcomes: No data - fewer recalls and fewer inadequate specimens with LBC may improve quality of lifeConvenience: no need to return for HPV testing in case of a positive result

Lower than comparators

Major improvement

0

1

2

3

I6

Public health interest

Screening programs for cervical cancer in South Africa: mortality decreased by 50% for white women, 40% for Asian women but rose for black women between 1960's 1990's. and Considerable interest to improve screening methods and uptake.

No risk reduction

Major risk reduction

0

1

2

3

I7

Type of medical service

Goal of intervention: improve outcome of the disease due to early detection; 5-years survival ranging between close to 100% for early stage and 5% to 15% for late stage

Minor Service

Major Service (e.g. cure)

Economics

E1

Budget impact on health plan

Intervention price: LBC - R180.00; CC - R103.10Incremental cost per pt screened with LBC: R76.90Annual projected budget impact per 1000 women: R76,900

Substantial additional spending

Substantial savings

0

1

2

3

E2

Cost-effectiveness of intervention

Cost per life-year gained: R764,000Cost per QALY gained: R758,000

Not cost-effective

Highly cost-effective

0

1

2

3

E3

Impact on other spending

Incremental colposcopy cost per patient over lifetime horizon: R 8

(excludes drug cost, see E1)No available data on other spending (including treatment of cancer and pre-cancer)

Substantial additional spending

Substantial Savings

0

1

2

3

Quality of evidence

Q1

Adherence to requirements of decisionmaking body

Not applicable for case study

Low adherence

High adherence

Q2

Completeness and consistency of reporting evidence

Quality score: Clinical data:75% - primary and secondary outcome measures as well as sensitivity analyses not clearly specified; Economic evaluation: 50% - disaggregated cost not reported; incomplete reporting of effectiveness outcomes

Many gaps/inconsistent

Complete & consistent

0

1

2

3

Q3

Relevance and validity of evidence

Quality score: Clinical data:75% -most relevant outcome not assessed (morbidity and mortality due to cervical cancer);Economic evaluation: 75% - Canadian screening coverage and HPV epidemiology not completely applicable for South African private payer setting

Low relevance/validity

High relevance/validity

0

1

2

3