Journal of Advanced Nursing, 1979, 4, 299-306
The relationship hetween the ievei of nurse staffing and the patterns of patient care and staff activity Brian Moores M.Sc. Ph.D.
Anne Moult* B.Sc. M.Sc. Department of Management Sciences, University of Manchester, Institute of Science, Manchester Acceptedfor publication 30 November 1978
MOORE B . AND MOULT A. (1979) Journal of Advanced Nursing 4, 299-306
The relationship between the level of nurse staffing and the patterns of patient care and staff activity Data relating to patterns of nursing activity and all the interactions occurring hetween patients and nursing staff were obtained during observation periods of i week's duration on a total of 20 wards in three hospitals. When analysed these data reveal that in all wards there is a strong relationship between the nursing hours available and the various components of the nursing care provided. Indeed the linear relationship is so strong in some cases that it appears that these wards are a long way short of the staffing levels which produced the 'saturation' effect described by several American authors. The analysis of the interactions reveals, as one would suspect, that a disproportionate number of patients account for a large proportion of the interactions, but two further analyses will, it is hoped, shed a little more hght on this aspect of the ward situation. The data are being subject to further examination to ascertain whether extra nursing staff leads to more or longer interactions and a longer term investigation is under way which is focused on patients with comparable diagnoses in different hospitals.
INTRODUCTION In a previous article (Moores & Moult 1979) the background and organization of this study were described in some detail as were the results pertaining to the overall patterns of nursing activity. Because the data collected could be associated with particular days and the availability of nurses varied not only between the three hospitals but also between different days on the same ward, it was possible to analyse the material in order to ascertain the nature of the relationship, if any, between the nursing care available and the patient care provided. In Figure i, for example, are shown the 63 values of these two measures for the basic care category of nursing on nine surgical wards, each point corresponding *Now with the Management Services Unit of the Metropolitan Police. 0309-2402/79/0300-0299 $02.00 © 1979 Blackwell Scientific Publications
299
Brian Moores and Anne Moult
300
to one ward on one day. The obvious close relationship between these two measures is reflected in the high product moment correlation coefficient of 0.92. Figure 2 is a similar plot for three orthopaedic wards but this time the nursing activities are those categorized as direct. This time the associated correlation coefficient is 0.94. Not all such coefficients are of this same order of magnitude as can be seen from Table i where the values for five categories of nursing care are set out, not otily for the eight medical, nine surgical and three orthopaedic wards considered together, but also for each ward separately. 150 r
r = 0-92
100XX
50 _LJL 2 3 Nursing Hours per patient FIGURE I
4
5
Relationship between Minutes of Basic care and Nursing Hours per patient on nine surgical wards 150
r = a94
Z" 100
m "S
X
X X
50
1
2
3
4
5
Nursing Hours per patient Relationship between Minutes of Direct care and Nursing Hours per patient on three orthopaedic wards
Nurse staffing, patient care and staff activity
301
TABLE I Coefficients of Correlation between the nursing hours per patient (NHPD) and the minutes of Basic Nursing, Technical Nursing, Direct Care and Indirect Care per patient for each day Basic Nursing
Technical Nursing
Direct Care
Indirect Care
003 105 109
0-66 0-82 0-76 o-8o 070
283
0-90
292 193
0-55 0-93 0-45 077
o-8i 0-97 0-97 078 091 079 0-49
0-94 0-97 0-99
IZ2
076 0-86 0-27 0-59 0-56 0-46 0-65 072 0-49 0-64
O'OO
0-90
0-37 o-8i
o-4r
0-90 o-o6 0-85 0-17
0-97 0-91 0-97 0-90
0-93 0-34 0-92 0-69
0-33 075 O'6o 0-38
0-89 o-8i 0-92 0-86
0-97 0-99 0-48 O'9i 0-92 0-81 0-92 0-72
0-38
0-82 095 079
0-96 0-98 o-6i
0-94 0-85
0-96
0-85
0-91
Speciality—Medical Ward 001
All Wards Average Speciality—Surg ical 004 006 104 106 110 222
223 232 243
All Wards Average
0-98 o-o8 0-87 0-97 0-92 0-32 0-77
077 0-58 0-84
0-97
0-93 0-90
079 0-81 0-93 0-87 091
0-94 0-44 0-97
Speciality—Orth opaedic
252
0-84 0-96 0-63
Ail Wards Average
0-95 0-81
0-19 0-42 0-29
All
0-82
0-51
005
242
031
0-85
Effects of more nursing hours The magnitude of most of these coefficients reveals what one would intuitively suspect, namely that the more nursing hours available on a particular day, the more nursing care patients receive. This would, initially, appear to suggest that in the United Kingdom the point has not yet been reached where additional nursing resources do not find their way to increased patient care. This fmding also serves to illustrate just how difficult is the problem of determining just what is an appropriate number of nursing staff in any particular situation. Correlation coefficients in this context could, however, be somewhat misleading in so far as one is ultimately interested in whether the additional nursing
Brian Moores and Anne Moult
302
time available is distributed in some different pattern. When the linear regression lines are determined for direct (DPC) and indirect patient care (IPC) for the nine surgical wards considered collectively the results are: DPC (mins/paticnt) = - 17.5+34-40 NHPD IPC (mins/patient) =4.2+25.72 NHPD Thus when the nursing hours per patient day figure (NHPD) is 2.5, the amount of direct patient care per patient is determined from the regression line to be ^8.5 minutes, and that for indirect patient care also coincidentally emerges as 68.5. In other words, at that level of staffing, when they are not involved on personal matters, the nurses devote half of their energies to activities wliich directly involve the patient and half to those which do not. When the availability of nurses is doubled, i.e. the NHPD figure is increased to five, the computed direct and indirect minutes of care turn out to be 154.5 and 132.8 respectively. Similar calculations can, of course, be performed for any NHPD value. When this is done for all three specialities, the results expressed in terms of proportion of nursing activity (excluding personal time) categorized as direct are depicted in Figure 3. 100 r
75
-~~~~
50
Surgical _ _prthopaedic ; . Medical
25
I
0
V
1-
1 NHPD
FIGURE 3 Percentage ofNursing Time devoted to Direct patient care for different Nursing Hours available (NHPD)
INTERACTIONS WITH
PATIENTS
It will be recalled from the previous article (Moores 8c Moult 1979) that information was also collected on all the interactions which occurred. It kad, originally, been the intention to examine the emerging interaction patterns for 'matched'
Nurse staffing, patient care and staff activity
303
samples of patients in the three institutions, but the number of similar cases was so few that this proved not to be a particularly revealing exercise. Nonetheless all the data collected had been processed through a computer and the printout organized around individual patients. Two such randomly selected printouts are displayed in Figures 4 and 5. The first relates to a female patient on an orthopaedic ward who had been admitted for a blade plate fixation, whilst the second relates to a male patient on a medical ward who had originally been admitted with myocardial infarction. (The activity code numbers referred to in these printouts correspond to those in the list of activities presented in the earher article).
PATIENT NO e
HOSP
3 SF-EC 3 UARP ZS2 SEX 2 BAt 7
BED NQ 9 HOUR NO 31
PATIENT NO 14 HOSP
3 SPEC 1 UARD ZV2 SEK 1 DAT 3
IlED NO a HOUR NO 21 NURSE
2
GRADE 9
ACTIVITY 113
HOU« NO 9
HDUR NO B MURSe 1* GRADE * NURSE 3 GRADE B NURSE 13 BRADE *
HOUR NO 10
HOUR NO
ACIIUITV 25 ACIIUITY 113 ACTIUtrr 25
CONVERSATION hEDICINE CONUERSflTION
HOUR HO 10 HOUR HO 13 NURSE 5 CRADE NUBSE 8 GRADE NURSE 7 GRAPE NURSE e GRADE
A ACIIUITV 2 ACTIVITT 2 ACTIUirr 2 ACTIVITY
11 5 10 10
BED PAHS UASHBOULS HOUTHUASHES PRESSURE AREAS PRESSURE AREAS
HOUR HO U HOUR HO 12 NURSE 5 GRADE
ACTIVITY 113
HG4fR NO 13 NURSE t? QRADE
5
*tTI«ITV
25
ACTIVITY
2
TPR CONVERSATION
HOUR NO 14 HOUR HO 15
HOUR NO IS
HOUR NO IA NURSE 3 GRAUE 8 NURSE ; GRODE 9
ACTIUIIY 113 ACTIVITY 25
HOUR NO 1 7 NURSE 7 GRADE 2 NURSE 3 ORADE S
HOUR ND It NURSE 6 GRADE 5 NURSE lA ORADE 4
tkCllVnt ACTIVITY
5 23
HOUR NQ 17
HOUR NO IB NURSE 4 QRADE 7 NURSE 3 GRADE 8 NURSE A GRADE 7
ACTIVITY ACTIVITY ACTIVITY
11 10 10
FIGURE 4
TIDYING BED
ACTIVITY 113 ACTIVITY 113 URIHALS
BED PANS PNESSURE AREAS PRESSURE AREAS
Individual pritttout of patient activity
COKUERSATtaN HOUR HO 1»
FIGURE 5
Individual printout of patient activity
Having such ready access to the data, it was a reasonably straight forw^ard matter to produce information relating to the number of interactions each patient experienced on any particular day and the number of nurses with whom the patient came into contact during these interactions. {The two are, of course, not necessarily the same due to some interactions involving more than a single nurse). A typical set of results for one ward is set out in Table 2 and when these data are displayed pictorially as in Figure 6, the shape, as one would expect, indicates that a small proportion of the patients account for a large proportion of the interactions. When the correlation coefficients between NHPD and the number of interactions per patient are determined, the values, which are set out in Table 3, are in almost all cases lower than those relating NHPD and the nursing activities.
304
Brian Moores and Anne
TABLE 2 The number of different nurses seen by each patient, number of different interactions received by each patient and the total number of interactions provided for each patient during the day for Ward io6
I
10
29
Total Number of Interactions 34
2
9
22
28
3
6
13
14
4
9
u
19
5
8
23
29
6
4
II
12
7
8
8
10
8
8
13
17
9
7
9
10
10
5
5
8
II
10
II
19
12
10
28
35
13
9
i6
l8
14
8
II
12
15
9
14
19
i6
5
7
7
Patient Number
Number of Different Number of Different Nurses Interactions
17
II
27
33
zS
7
14
i6
19
4
3
4
20
6
10
15
21
4
5
7
22
5
xo
12
23
6
II
14
24
10
29
38
25
9
28
30
26
I
I
I
27
7
10
II
28
5
9
10
29
7
13
15
30
3
4
4
Moult
Nurse staffing, patient care and staff activity
305
100
10 FIGURE 6
15 Number of patients
20
25
30
Percentage of interactions received by number of patients in one surgical ward arranged in descending order
Data Analysis One rather detailed analysis of the data produced considerable insight into the form that the additional direct case takes. It is often argued that if extra nurses were available this additional resource would find its way to mure things being done for the patients. On the other hand it could be that the time spent performing each task is increased but the number of tasks undertaken for each patient remain the same. To examine this question recourse was made to partial correlation and regression analysis. Consider, for example, the data relating to ward 003. From Tables i and 3 it will be seen that the normal product moment correlation coefficient between NHPD and the DPC/patient is 0.97, and that between the NHPD and the number of interactions per patient is 0.07. If the DPC figure is divided by that for the number of interactions, the time per interaction can be determined on each of the seven study days. The product moment correlation coefficient between this computed measure and the corresponding NHPD figures is 0.34. The partial correlation coefficient measures the degree of association between two variables with one or more others held constant. Li other words, it represents an attempt to demonstrate the impact of holding one or more variables constant when it is not possible to control for these variations in some experimentally controlled conditions. Thus, in this particular example, the partial correlation coefficient between NHPD and the number of interactions when the time per interaction is 'held' constant is 0.97, whilst that between NHPD and time per interaction with the number of interactions held constant is 0.98. Not all such
Brian Moores aud Anne Moult TABLE 3 Coefficients of Correlation and Partial Correlations between the tiursing hours per patient (NHPD) and the minutes of Direct care per patient, the number of interactions per patient and the number of minutes per interaction
I—NHPD; 2—Direct niinutes/Patient; 3—Number of interactions/Patient; 4—Direct minutes/Interaction ••13
'"I4-3
''13
Speciality—Medical 001
003 105 109 112
283 292
293
Ail Wards
o-8i 0-97 0-97 0-31 0-91 0-79 0-49 O-77 0-58
0-09 0-07 -0-31 0-42 0-13
0-90 0-41
0-75
0-55 -0-44 0-75 0-22
0*71 0-34 0-70 -031 075 0-21 0-59 0-14 0-29
o-6i 0-97 o-8i 0-31 0-83 0-52 — o-io 0-86 0-53
083 0-98 0-90 o-o8 093
— o-to 0-43 0-65 0-56
Speciality—Surgical 004
006 104
0.97
106
I-OO
no
0-90 0-96 0-89 0-89 o-8i 0-92
222
223 232 243
All Wards
Speciality—Orthopaedic 0-82 005 242 0-95 252
All Wards
0-79 0-94
0-39 0 36 0-73 0-62 o-6o 0-07 -0-42 0-21
0-03 0-28 0-85 0-48 O-2O
O-34 0-48 0-96 0-62
0-44
0-73
— 0-62 0-88 o-ii 0-6i
0-84 0-87 0-52 0-87
0-89 0-30 0-96 0-99 0-89 0-98 0-75 0-82 0-68 o-8i
0-32 083
-0-33 0-70
0*72 0
0-99 O-97 o-8o 094 0-82 I-00 o-8i 0-90
076 078 0-58 0-89
computed values were this high as is evident from Table 3, but they are sufficiently large to suggest that a more detailed analysis this time using regression methodology could well reveal whether the additional nursing time is more likely to impact upon the frequency of interactions than upon the duration of these interactions. Such an atialysis is, in fact, currently heing pursued and should it reveal anything of any great consequence the results will be reported at some future date. Reference MOORES B . & MOULT A. (1979) Patterns of nurse activity. Journal of Advanced Nursing 4, i37-t49.
The relationship hetween the ievei of nurse staffing and the patterns of patient care and staff activity Brian Moores M.Sc. Ph.D.
Anne Moult* B.Sc. M.Sc. Department of Management Sciences, University of Manchester, Institute of Science, Manchester Acceptedfor publication 30 November 1978
MOORE B . AND MOULT A. (1979) Journal of Advanced Nursing 4, 299-306
The relationship between the level of nurse staffing and the patterns of patient care and staff activity Data relating to patterns of nursing activity and all the interactions occurring hetween patients and nursing staff were obtained during observation periods of i week's duration on a total of 20 wards in three hospitals. When analysed these data reveal that in all wards there is a strong relationship between the nursing hours available and the various components of the nursing care provided. Indeed the linear relationship is so strong in some cases that it appears that these wards are a long way short of the staffing levels which produced the 'saturation' effect described by several American authors. The analysis of the interactions reveals, as one would suspect, that a disproportionate number of patients account for a large proportion of the interactions, but two further analyses will, it is hoped, shed a little more hght on this aspect of the ward situation. The data are being subject to further examination to ascertain whether extra nursing staff leads to more or longer interactions and a longer term investigation is under way which is focused on patients with comparable diagnoses in different hospitals.
INTRODUCTION In a previous article (Moores & Moult 1979) the background and organization of this study were described in some detail as were the results pertaining to the overall patterns of nursing activity. Because the data collected could be associated with particular days and the availability of nurses varied not only between the three hospitals but also between different days on the same ward, it was possible to analyse the material in order to ascertain the nature of the relationship, if any, between the nursing care available and the patient care provided. In Figure i, for example, are shown the 63 values of these two measures for the basic care category of nursing on nine surgical wards, each point corresponding *Now with the Management Services Unit of the Metropolitan Police. 0309-2402/79/0300-0299 $02.00 © 1979 Blackwell Scientific Publications
299
Brian Moores and Anne Moult
300
to one ward on one day. The obvious close relationship between these two measures is reflected in the high product moment correlation coefficient of 0.92. Figure 2 is a similar plot for three orthopaedic wards but this time the nursing activities are those categorized as direct. This time the associated correlation coefficient is 0.94. Not all such coefficients are of this same order of magnitude as can be seen from Table i where the values for five categories of nursing care are set out, not otily for the eight medical, nine surgical and three orthopaedic wards considered together, but also for each ward separately. 150 r
r = 0-92
100XX
50 _LJL 2 3 Nursing Hours per patient FIGURE I
4
5
Relationship between Minutes of Basic care and Nursing Hours per patient on nine surgical wards 150
r = a94
Z" 100
m "S
X
X X
50
1
2
3
4
5
Nursing Hours per patient Relationship between Minutes of Direct care and Nursing Hours per patient on three orthopaedic wards
Nurse staffing, patient care and staff activity
301
TABLE I Coefficients of Correlation between the nursing hours per patient (NHPD) and the minutes of Basic Nursing, Technical Nursing, Direct Care and Indirect Care per patient for each day Basic Nursing
Technical Nursing
Direct Care
Indirect Care
003 105 109
0-66 0-82 0-76 o-8o 070
283
0-90
292 193
0-55 0-93 0-45 077
o-8i 0-97 0-97 078 091 079 0-49
0-94 0-97 0-99
IZ2
076 0-86 0-27 0-59 0-56 0-46 0-65 072 0-49 0-64
O'OO
0-90
0-37 o-8i
o-4r
0-90 o-o6 0-85 0-17
0-97 0-91 0-97 0-90
0-93 0-34 0-92 0-69
0-33 075 O'6o 0-38
0-89 o-8i 0-92 0-86
0-97 0-99 0-48 O'9i 0-92 0-81 0-92 0-72
0-38
0-82 095 079
0-96 0-98 o-6i
0-94 0-85
0-96
0-85
0-91
Speciality—Medical Ward 001
All Wards Average Speciality—Surg ical 004 006 104 106 110 222
223 232 243
All Wards Average
0-98 o-o8 0-87 0-97 0-92 0-32 0-77
077 0-58 0-84
0-97
0-93 0-90
079 0-81 0-93 0-87 091
0-94 0-44 0-97
Speciality—Orth opaedic
252
0-84 0-96 0-63
Ail Wards Average
0-95 0-81
0-19 0-42 0-29
All
0-82
0-51
005
242
031
0-85
Effects of more nursing hours The magnitude of most of these coefficients reveals what one would intuitively suspect, namely that the more nursing hours available on a particular day, the more nursing care patients receive. This would, initially, appear to suggest that in the United Kingdom the point has not yet been reached where additional nursing resources do not find their way to increased patient care. This fmding also serves to illustrate just how difficult is the problem of determining just what is an appropriate number of nursing staff in any particular situation. Correlation coefficients in this context could, however, be somewhat misleading in so far as one is ultimately interested in whether the additional nursing
Brian Moores and Anne Moult
302
time available is distributed in some different pattern. When the linear regression lines are determined for direct (DPC) and indirect patient care (IPC) for the nine surgical wards considered collectively the results are: DPC (mins/paticnt) = - 17.5+34-40 NHPD IPC (mins/patient) =4.2+25.72 NHPD Thus when the nursing hours per patient day figure (NHPD) is 2.5, the amount of direct patient care per patient is determined from the regression line to be ^8.5 minutes, and that for indirect patient care also coincidentally emerges as 68.5. In other words, at that level of staffing, when they are not involved on personal matters, the nurses devote half of their energies to activities wliich directly involve the patient and half to those which do not. When the availability of nurses is doubled, i.e. the NHPD figure is increased to five, the computed direct and indirect minutes of care turn out to be 154.5 and 132.8 respectively. Similar calculations can, of course, be performed for any NHPD value. When this is done for all three specialities, the results expressed in terms of proportion of nursing activity (excluding personal time) categorized as direct are depicted in Figure 3. 100 r
75
-~~~~
50
Surgical _ _prthopaedic ; . Medical
25
I
0
V
1-
1 NHPD
FIGURE 3 Percentage ofNursing Time devoted to Direct patient care for different Nursing Hours available (NHPD)
INTERACTIONS WITH
PATIENTS
It will be recalled from the previous article (Moores 8c Moult 1979) that information was also collected on all the interactions which occurred. It kad, originally, been the intention to examine the emerging interaction patterns for 'matched'
Nurse staffing, patient care and staff activity
303
samples of patients in the three institutions, but the number of similar cases was so few that this proved not to be a particularly revealing exercise. Nonetheless all the data collected had been processed through a computer and the printout organized around individual patients. Two such randomly selected printouts are displayed in Figures 4 and 5. The first relates to a female patient on an orthopaedic ward who had been admitted for a blade plate fixation, whilst the second relates to a male patient on a medical ward who had originally been admitted with myocardial infarction. (The activity code numbers referred to in these printouts correspond to those in the list of activities presented in the earher article).
PATIENT NO e
HOSP
3 SF-EC 3 UARP ZS2 SEX 2 BAt 7
BED NQ 9 HOUR NO 31
PATIENT NO 14 HOSP
3 SPEC 1 UARD ZV2 SEK 1 DAT 3
IlED NO a HOUR NO 21 NURSE
2
GRADE 9
ACTIVITY 113
HOU« NO 9
HDUR NO B MURSe 1* GRADE * NURSE 3 GRADE B NURSE 13 BRADE *
HOUR NO 10
HOUR NO
ACIIUITV 25 ACIIUITY 113 ACTIUtrr 25
CONVERSATION hEDICINE CONUERSflTION
HOUR HO 10 HOUR HO 13 NURSE 5 CRADE NUBSE 8 GRADE NURSE 7 GRAPE NURSE e GRADE
A ACIIUITV 2 ACTIVITT 2 ACTIUirr 2 ACTIVITY
11 5 10 10
BED PAHS UASHBOULS HOUTHUASHES PRESSURE AREAS PRESSURE AREAS
HOUR HO U HOUR HO 12 NURSE 5 GRADE
ACTIVITY 113
HG4fR NO 13 NURSE t? QRADE
5
*tTI«ITV
25
ACTIVITY
2
TPR CONVERSATION
HOUR NO 14 HOUR HO 15
HOUR NO IS
HOUR NO IA NURSE 3 GRAUE 8 NURSE ; GRODE 9
ACTIUIIY 113 ACTIVITY 25
HOUR NO 1 7 NURSE 7 GRADE 2 NURSE 3 ORADE S
HOUR ND It NURSE 6 GRADE 5 NURSE lA ORADE 4
tkCllVnt ACTIVITY
5 23
HOUR NQ 17
HOUR NO IB NURSE 4 QRADE 7 NURSE 3 GRADE 8 NURSE A GRADE 7
ACTIVITY ACTIVITY ACTIVITY
11 10 10
FIGURE 4
TIDYING BED
ACTIVITY 113 ACTIVITY 113 URIHALS
BED PANS PNESSURE AREAS PRESSURE AREAS
Individual pritttout of patient activity
COKUERSATtaN HOUR HO 1»
FIGURE 5
Individual printout of patient activity
Having such ready access to the data, it was a reasonably straight forw^ard matter to produce information relating to the number of interactions each patient experienced on any particular day and the number of nurses with whom the patient came into contact during these interactions. {The two are, of course, not necessarily the same due to some interactions involving more than a single nurse). A typical set of results for one ward is set out in Table 2 and when these data are displayed pictorially as in Figure 6, the shape, as one would expect, indicates that a small proportion of the patients account for a large proportion of the interactions. When the correlation coefficients between NHPD and the number of interactions per patient are determined, the values, which are set out in Table 3, are in almost all cases lower than those relating NHPD and the nursing activities.
304
Brian Moores and Anne
TABLE 2 The number of different nurses seen by each patient, number of different interactions received by each patient and the total number of interactions provided for each patient during the day for Ward io6
I
10
29
Total Number of Interactions 34
2
9
22
28
3
6
13
14
4
9
u
19
5
8
23
29
6
4
II
12
7
8
8
10
8
8
13
17
9
7
9
10
10
5
5
8
II
10
II
19
12
10
28
35
13
9
i6
l8
14
8
II
12
15
9
14
19
i6
5
7
7
Patient Number
Number of Different Number of Different Nurses Interactions
17
II
27
33
zS
7
14
i6
19
4
3
4
20
6
10
15
21
4
5
7
22
5
xo
12
23
6
II
14
24
10
29
38
25
9
28
30
26
I
I
I
27
7
10
II
28
5
9
10
29
7
13
15
30
3
4
4
Moult
Nurse staffing, patient care and staff activity
305
100
10 FIGURE 6
15 Number of patients
20
25
30
Percentage of interactions received by number of patients in one surgical ward arranged in descending order
Data Analysis One rather detailed analysis of the data produced considerable insight into the form that the additional direct case takes. It is often argued that if extra nurses were available this additional resource would find its way to mure things being done for the patients. On the other hand it could be that the time spent performing each task is increased but the number of tasks undertaken for each patient remain the same. To examine this question recourse was made to partial correlation and regression analysis. Consider, for example, the data relating to ward 003. From Tables i and 3 it will be seen that the normal product moment correlation coefficient between NHPD and the DPC/patient is 0.97, and that between the NHPD and the number of interactions per patient is 0.07. If the DPC figure is divided by that for the number of interactions, the time per interaction can be determined on each of the seven study days. The product moment correlation coefficient between this computed measure and the corresponding NHPD figures is 0.34. The partial correlation coefficient measures the degree of association between two variables with one or more others held constant. Li other words, it represents an attempt to demonstrate the impact of holding one or more variables constant when it is not possible to control for these variations in some experimentally controlled conditions. Thus, in this particular example, the partial correlation coefficient between NHPD and the number of interactions when the time per interaction is 'held' constant is 0.97, whilst that between NHPD and time per interaction with the number of interactions held constant is 0.98. Not all such
Brian Moores aud Anne Moult TABLE 3 Coefficients of Correlation and Partial Correlations between the tiursing hours per patient (NHPD) and the minutes of Direct care per patient, the number of interactions per patient and the number of minutes per interaction
I—NHPD; 2—Direct niinutes/Patient; 3—Number of interactions/Patient; 4—Direct minutes/Interaction ••13
'"I4-3
''13
Speciality—Medical 001
003 105 109 112
283 292
293
Ail Wards
o-8i 0-97 0-97 0-31 0-91 0-79 0-49 O-77 0-58
0-09 0-07 -0-31 0-42 0-13
0-90 0-41
0-75
0-55 -0-44 0-75 0-22
0*71 0-34 0-70 -031 075 0-21 0-59 0-14 0-29
o-6i 0-97 o-8i 0-31 0-83 0-52 — o-io 0-86 0-53
083 0-98 0-90 o-o8 093
— o-to 0-43 0-65 0-56
Speciality—Surgical 004
006 104
0.97
106
I-OO
no
0-90 0-96 0-89 0-89 o-8i 0-92
222
223 232 243
All Wards
Speciality—Orthopaedic 0-82 005 242 0-95 252
All Wards
0-79 0-94
0-39 0 36 0-73 0-62 o-6o 0-07 -0-42 0-21
0-03 0-28 0-85 0-48 O-2O
O-34 0-48 0-96 0-62
0-44
0-73
— 0-62 0-88 o-ii 0-6i
0-84 0-87 0-52 0-87
0-89 0-30 0-96 0-99 0-89 0-98 0-75 0-82 0-68 o-8i
0-32 083
-0-33 0-70
0*72 0
0-99 O-97 o-8o 094 0-82 I-00 o-8i 0-90
076 078 0-58 0-89
computed values were this high as is evident from Table 3, but they are sufficiently large to suggest that a more detailed analysis this time using regression methodology could well reveal whether the additional nursing time is more likely to impact upon the frequency of interactions than upon the duration of these interactions. Such an atialysis is, in fact, currently heing pursued and should it reveal anything of any great consequence the results will be reported at some future date. Reference MOORES B . & MOULT A. (1979) Patterns of nurse activity. Journal of Advanced Nursing 4, i37-t49.
