Automated statistical tracker

Claim 3

The method of claim 1, wherein: (a) said messages generated by said system are of a plurality of types, predefined limits being established for each of said message types; (b) said counting step comprises forming message subgroups for each of said plurality of message types, and (c) said comparing step comprises comparing the number of messages in each of said subgroups of each message type with said corresponding predefined limits.

Claim 4

The method of claim 2, wherein said system is a clinical event monitor comprising a plurality of medical logic modules, each of said plurality of medical logic modules generating a different one of said plurality of types of messages.

Claim 5

The method of claim 3, wherein said statistically significant time periods each consist of one day.

Claim 6

The method of claim 1, wherein said predefined limits include an upper probability limit and a lower probability limit.

Claim 7

The method of claim 5, wherein said upper probability limit represents the highest statistically acceptable value for said number of messages, and said lower probability limit represents the lowest statistically acceptable value for said number of messages.

Claim 8

The method of claim 6, wherein said upper probability limit and said lower probability limit are derived from a statistically significant sample of past message data.

Claim 9

The method of claim 7, further including the steps of: (e) retrieving said sample of past message data from storage; and (f) generating said upper and said lower probability limits from said retrieved sample of past message data prior to step (c).

Claim 10

The method of claim 8, wherein said upper probability limit is determined from an algorithm based on poisson and normal distributions applied to said sample of past message data.

Claim 11

The method of claim 9, wherein said poisson distribution has a single probability threshold of 0.001, said normal distribution is based on the z statistic with a single probability threshold of 0.001, and said upper algorithm is based on a maximum value determined from said poisson and normal distributions.

Claim 12

The method of claim 8, wherein said lower probability limit is determined from an algorithm based on a normal distribution and a binomial distribution applied to said sample of past message data.

Claim 13

The method of claim 11, wherein said binomial distribution comprises the consecutive zero algorithm.

Claim 14

The method of claim 12, wherein said normal distribution is based on the z statistic with a single probability threshold of 0.001, said consecutive zero algorithm has a single probability threshold of 0.001, and said lower algorithm is based on a maximum value determined from said normal distribution and said consecutive zero algorithm.

Claim 15

The method of claim 8, further including the step of updating said stored sample of past message data based upon said messages collected in step (b).

Claim 16

The method of claim 14, wherein said statistically significant time period is one day, and said past message data comprises data from a preselected number of days.

Claim 17

The method of claim 8, wherein said statistically unusual number of messages is detected whenever said number of messages is greater than said upper probability limit or is less than said lower probability limit.

Claim 18

The method of claim 16, wherein said alert signal generated in step (d) is electronic mail sent to a system administrator.

Claim 19

Apparatus for automatically monitoring a system to detecting malfunctions, wherein said system generates messages based upon selected data, comprising: (a) input means for receiving messages generated by said system; (b) counting means, coupled to said input means, for collecting said messages received by said input means over statistically significant time periods to form message subgroups including selected numbers of messages; (c) comparator means, coupled to said counting means, for receiving the subgroups from said counting means and comparing said number of messages in each of said subgroups with predefined limits for that subgroup to determine whether said number of messages in that subgroup is statistically unusual; (d) alarm means, coupled to said comparator means, for generating an alert signal whenever a statistically unusual number of messages is detected by said comparator means.

Claim 20

Apparatus for automatically monitoring a system to detecting malfunctions, wherein said system generates messages based upon selected data, comprising: (a) input means for receiving messages generated by said system; (b) counting means, coupled to said input means, for collecting said messages received by said input means over statistically significant time periods to form message subgroups including selected numbers of messages; (c) comparator means, coupled to said counting means, for receiving the subgroups from said counting means and comparing said number of messages in each of said subgroups with predefined limits for that subgroup to determine whether said number of messages in that subgroup is statistically unusual; (d) alarm means, coupled to said comparator means, for generating an alert signal whenever a statistically unusual number of messages is detected by said comparator means.

Claim 21

The apparatus of claim 18, wherein: (a) said messages generated by said system are of a plurality of types, predefined limits being established for each of said message types; (b) said counting means forms message subgroups for each of said plurality of message types, and (c) said comparator means compares the number of messages in each of said subgroups of each message type with said corresponding predefined limits.

Claim 22

The apparatus of claim 19, wherein said system is a clinical event monitor comprising a plurality of medical logic modules, each of said plurality of medical logic modules generating a different one of said plurality of types of messages.

Claim 23

The apparatus of claim 20, wherein said statistically significant time periods each consist of one day.

Claim 24

The apparatus of claim 18, wherein said predefined limits include an upper probability limit and a lower probability limit.

Claim 25

The apparatus of claim 22, wherein said upper probability limit represents the highest statistically acceptable value for said number of messages, and said lower probability limit represents the lowest statistically acceptable value for said number of messages.

Claim 26

The apparatus of claim 23, wherein said upper probability limit and said lower probability limit are derived from a statistically significant sample of past message data.

Claim 27

The apparatus of claim 24, further including: (a) limit generating means, coupled to said comparator means, for generating said upper and said lower probability limits from said sample of past message data; (b) data storage means, coupled to said limit generating means, for storing said sample of past message data, said limit generating means retrieving said sample of past message data from said data storage means prior to generating said probability limits.

Claim 28

The apparatus of claim 25, wherein said upper probability limit is determined from an algorithm based on poisson and normal distributions applied to said sample of past message data.

Claim 29

The apparatus of claim 26, wherein said poisson distribution has a single probability threshold of 0.001, said normal distribution is based on the z statistic with a single probability threshold of 0.001, and said upper algorithm is based on a maximum value determined from said poisson and normal distributions.

Claim 30

The apparatus of claim 25, wherein said lower probability limit is determined from an algorithm based on a normal distribution and a binomial distribution applied to said sample of past message data.

Claim 31

The apparatus of claim 28, wherein said binomial distribution comprises the consecutive zero algorithm.

Claim 32

The apparatus of claim 29, wherein said normal distribution is based on the z statistic with a single probability threshold of 0.001, said consecutive zero algorithm has a single probability threshold of 0.001, and said lower algorithm is based on a maximum value determined from said normal distribution and said consecutive zero algorithm.

Claim 33

The apparatus of claim 25, further including data updating means, coupled to said comparator means and said data storage means, for revising said stored sample of past message data to include said messages collected by said counting means.

Claim 34

The apparatus of claim 31, wherein said statistically significant time period is one day, and said past message data comprises data from a preselected number of days.

Claim 35

The apparatus of claim 25, wherein said statistically unusual number of messages is detected by said comparator means whenever said number of messages is greater than said upper probability limit or is less than said lower probability limit.

Claim 36

The apparatus of claim 33, wherein said alarm means includes electronic mail sent to a system administrator.