Not for Sale

• #### Index 453-458

because a vaccine must be accepted by both provider and patient to be utilized. Thus, even the highest degree of acceptance by the patient is cancelled by near zero acceptance by the provider, and vice versa. This logic supports the multiplicative form of aggregation of provider and lay scores. All methods of aggregation have been carried forward, however, to illustrate that they ultimately give similar results. Data shown in Table 6.4 could be used to generate a ranking of vaccine candidates on the basis of acceptance scores.

#### Translation of Acceptance Scores into Anticipated Use Rates

To calculate benefits that might accrue from developing each vaccine candidate, it is necessary to calculate an anticipated use rate, it is not obvious, however, how the acceptance “scores” (Table 6.4) map onto the range of utilization “rates” from 0 to 100 percent. Two steps were followed to derive such a mapping.

First, to relate acceptance scores to likely actual use rates, the HBM scoring system was applied “retrospectively” to the introduction of certain vaccines now in use: the influenza, pneumococcal,* and pertussis vaccines. Relatively reliable information was available on voluntary use rates for these vaccines. Pertussis had a utilization rate of approximately 85 percent prior to the legal requirement of vaccination for school entry; influenza, 20 percent; and pneumococcal, 10 percent within the high-risk target populations (Orenstein, personal communication, 1983). HBM scores were estimated retrospectively for these vaccines, adopting what were thought to be the perspectives of appropriate lay and provider groups at a time prior to the introduction of the vaccines. These scores are shown in Tables 6.5 and 6.6.

Next, a logistic regression was performed to define the mathematical relationship between the “retrospective” acceptance scores in our HBM model and the observed use rates. The data points were the use rate-combined score combinations for pertussis, influenza, and pneumococcal vaccines shown in Table 6.6.

The regression equations were as follows, where P=use rate, and S=acceptance score:

 (Case A: lay and provider weighted additive scores combined in additive fashion) (Case B: lay and provider weighted additive scores combined in multiplicative fashion) (Cases C&D: weighted multiplicative scores combined in either additive or multiplicative fashion)
 * Streptococcus pneumoniae.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001