3

Determining Priorities
in Monitoring
40 AN INTRODUCTION TO QUALITY ASSURANCE IN HEALTH CARE

A
nyone setting out to monitor the quality of care is likely to be
confronted with many things vying for attention; so many that one
may not know where to begin. It is important, therefore, to have
some guidelines for selection. In this section of my exploration, I shall
try to offer such a guide. For convenience of discussion, I shall organize
it as follows: importance to patient welfare . . . additional institutional
interests . . . feasibility. These I shall take up in order.

Importance to Patient Welfare

The rule that governs importance to patient welfare may be summarized

by the expression "maximum achievable benefit," a guideline introduced
by John Williamson. Accordingly, the characteristics of a problem that
determine its importance are as follows:12-13

1. The problem is believed to occur frequently.

2. Error or failure in performance is known or believed to occur
frequently.
3. When it occurs, such error or failure in performance is believed
or known to have serious consequences to health, and is costly.
4. The error or failure in question can be rather easily corrected.

I know of no generally acceptable formula that assigns weights to these

several properties so that they can be summarized by a single index of
importance. That is a matter for informed judgment. Briefly, one could
say: select what is frequent, grievous, and correctable.

Additional Institutional Interests

Because the welfare of patients is the primary responsibility of providers,

whether these are institutions or individuals, "importance," as described
above, continues to be the paramount concern of all providers. That said,
one should be aware that other considerations can also influence insti-
Determining Priorities in Monitoring 41

tutional decisions on what requires more urgent attention. I list and dis-
cuss some of these as follows.

Meeting Externally Imposed Requirements

By virtue of their legal and administrative consequences, such require-
ments acquire almost automatic priority even when, if the requirement
did not exist, the institution would have preferred to direct its attention
elsewhere.

Risk Reduction
There is, of course, a significant degree of correlation between the adverse
experiences of patients and consequent risk to the organization. Never-
theless, in assessing risk, attention focuses mainly on whether or not the
institution is likely to be sued, attract the unfavorable attention of regu-
lators or insurers, or lose its good reputation. In countries where litigious-
ness is rampant, the prevention of being sued is the major consideration
in risk management. If a suit is successful, large penalties may be assessed.
And even if unsuccessful, a suit is costly in legal fees and other expenses,
raises the cost of insurance against malpractice (actual or alleged), and
damages the institution's reputation.

Institutional Enhancement
The reduction of risk contributes to patient welfare, reduces cost to the
institution, and protects its reputation. In a previous section I also pointed
out that cost reduction could be attained by the exclusion of harmful or
"useless" care, a goal quality monitoring should contribute to.
The reduction of risk and cost, as well as the improvement of quality,
enhance the reputation of an institution and attract patients, practitioners,
and other resources to it. The very fact that an institution demonstrates
its dedication to quality by having established a mechanism to watch over
it, if made known to the public, recommends the institution to those who
may wish to use it.
42 AN INTRODUCTION TO QUALITY ASSURANCE IN HEALTH CARE

Representativeness
It is not often that truly representative sampling, in the statistical sense,
is used in quality monitoring. Reasonable precautions against bias, for
example in selecting medical records for review, are generally sufficient.
Perhaps this is because quality monitoring is not perceived as research. It
is, ordinarily, a much less rigorous effort to collect information that reveals
the kinds of care that need attention in a specific setting. Precise quan-
tification of problems is not needed, nor is there any effort to generalize.
With that having been said, monitoring systems should have in mind
all the clinical activities of an institution and attempt to include all these
as actual or potential subjects for performance assessment. This calls for
some scheme of what may be called illustrative sampling, of which the
tracer method already described could be one example.

Fairness
A plan for quality reconnaissance should also be alert to yet another prin-
ciple: that of fairness. I mean that monitoring should not be concentrated
so heavily on some units, functions, or individuals to be seen as unduly
biased or bordering on the persecutory, unless there is clear reason for
such concentration of attention. One advantage of aiming for represen-
tativeness is to avoid the perception of unfair targeting.

Feasibility

Not every kind of monitoring one might wish to institute is feasible or

practicable. There could be a number of obstacles or, on the contrary,
facilitators, that influence when monitoring begins and what directions it
might take. These include organizational readiness and operational feas-
ibility.

Organizational readiness. It is best, when one considers where and

what to monitor, to search for units in which there is a perception that
quality needs to be improved, a desire to improve, and a willingness to
Determining Priorities in Monitoring 43

institute monitoring. Often this will depend on the presence within a unit
(or the institution as a whole) of one or more respected leaders who are
willing to cooperate. It is all the better if they themselves request that
monitoring be instituted.

Operational requirements. A number of "operational" or "techni-

cal" requirements can also influence the way one begins and continues
to monitor. What one chooses to assess should lend itself to valid judg-
ments of quality because the science of health care is so well developed
that one knows what is good or not good to do or accomplish. This is
reflected, as we shall see, in one's ability to formulate valid criteria of
quality. Moreover, there should be accurate and rather easily available
information concerning the activities and results of care. And if the or-
ganization desires to judge primarily the performance of the practitioners
themselves, the care implemented and the results obtained should not
depend too much on the cooperation of the patient. In other situations,
of course, the degree of patient cooperation would be important to in-
clude, or it could be the major focus of attention.