Health Care Management Science 3 (2000) 237–247

237

The relationship between market orientation and performance in the hospital industry: A structural equations modeling approach P.S. Raju a,∗ , Subhash C. Lonial a , Yash P. Gupta b and Craig Ziegler c a

College of Business and Public Administration, University of Louisville, Louisville, KY 40292, USA b Business School, University of Washington, Seattle, WA 98195, USA c Academic Consultant, University of Louisville, KY, USA

Received 22 April 1999; accepted 4 February 2000

There is general consensus in the research literature that market orientation is related to organizational performance. This study examines this relationship in the hospital industry. One unique feature of this study is that both market orientation and performance are conceptualized as being multi-dimensional constructs. Hence the technique of Structural Equations Modeling (SEM) is used to examine the relationship. Analyses were based on market orientation and performance data obtained from 175 hospitals in a five-state region of the United States. The SEM results confirm the multi-dimensional nature of both market orientation and performance, and the strong relationship between the constructs. Interestingly, this relationship is found to be much stronger for smaller hospitals than for larger hospitals. Implications for the hospital industry are discussed. Keywords: market orientation, hospital performance, structural equations modeling, hospital size, customer orientation, customer service, internal quality, market development, responsiveness to customers, responsiveness to competition

1. Introduction Several studies in recent years have confirmed that the market orientation of firms is related to their performance [1–3]. In this paper we examine this relationship within the hospital industry. It is well known that hospitals have been quite slow in embracing marketing related concepts. Hospitals have often been known to incorrectly equate marketing orientation with the use of public relations and advertising. While these functions are quite important to hospitals, recent studies have shown that true marketing orientation encompasses several dimensions that go beyond these functions [1,4]. It is therefore very important for hospitals to understand the true nature of market orientation in order to survive in today’s intensely competitive environment. The other side of the equation is performance. Again, performance is often mistakenly perceived as meaning only financial performance. In this paper we advance the notion that performance is also a multidimensional concept. Since both market orientation and performance are conceptualized as multidimensional concepts the examination of the relationship between these concepts in the hospital industry raises many interesting issues. Some of the major issues are: 1. What is the dimensionality of market orientation? 2. What is the dimensionality of performance? 3. What is the nature of the relationship between market orientation and performance? 4. What are the implications of this relationship for hospitals and for researchers? ∗

Corresponding author. E-mail: [email protected].

 Baltzer Science Publishers BV

In this paper we hope to shed some light on these issues. We first briefly discuss the theoretical background for the study, namely the literature relating to market orientation, performance and the relationship between these two concepts. The methodology and data analyses are discussed next. A structural equations modeling (SEM) technique is used to examine the dimensionality of market orientation and performance and to examine the nature of the relationship between the two concepts. The paper concludes with the discussion of the results and the implications of these results for the hospital industry and for researchers.

2. Theoretical background In the past market orientation has been defined in several different ways. In one of the pioneering articles on the subject of market orientation Shapiro [5] states, “After years of research, I’m convinced that the term “market oriented” represents a set of processes touching on all aspects of the company. It’s a great deal more than the clich´e “getting close to the customer” ”. Kotler and Clarke [6] define market orientation as a tendency to “determine the needs and wants of target markets and to satisfy them through the design, communication, pricing, and delivery of appropriate and competitively viable products and services”. According to them the five major attributes that characterize market orientation are a customer philosophy, integrated marketing organization, adequate marketing information, strategic orientation, and operational efficiency. According to Narver and Slater [1], the driving force behind market orientation is the desire to create superior value

238

P.S. Raju et al. / Market orientation and performance in the hospital industry

for customers and attain sustainable competitive advantage. They consider market orientation to comprise three major dimensions, namely customer orientation, competitor orientation, and interfunctional coordination. Based on in-depth interviews with 62 managers in four US cities, Kohli and Jaworski [4] define market orientation as, “the organizationwide generation of market intelligence pertaining to current and future customer needs, dissemination of the intelligence across departments, and organizationwide responsiveness to it”. Therefore they consider the three dimensions of market orientation to be intelligence generation, intelligence dissemination, and responsiveness. It is clear from these definitions that market orientation is perceived to be a multidimensional concept although researchers do not seem to agree on the exact number or nature of the dimensions. From a theoretical perspective, we believe that both the Narver and Slater [1] and the Kohli and Jaworski [4] frameworks have considerable validity. These researchers have conducted a number of studies that support their respective frameworks and both frameworks are intuitively appealing with respect to the nature of the dimensions. For purposes of the present research we have modified the MARKOR instrument designed by Kohli et al. [7] to measure market orientation in a healthcare context. In an earlier study, we used exploratory factor analysis to examine the factor structure of market orientation and found that the factor structure did have similarities to both frameworks [12]. We discuss this in more detail later in this paper and provide further validation of the factor structure using the technique of confirmatory factor analysis. The other major construct in this study is hospital performance. As the hospital industry becomes more competitive this construct has received a considerable degree of attention in the healthcare literature [8–10]. These researchers have generally examined hospital performance in the context of total quality management (TQM). Performance in this context is often perceived in terms of both effectiveness and efficiency. Effectiveness relates to whether the hospital systematically provides higher quality and more appropriate services and efficiency relates to lowering the cost of providing these improved services. Counte et al. [9] state that efficiency can be assessed in three areas – financial, operations, and human resources – and effectiveness can be assessed in four areas – financial, operations, human resources, and market. While business executives often look at organizational performance purely from the perspective of financial outcomes, the research in the healthcare area shows that this could be quite short sighted. In this paper, we therefore conceptualize hospital performance as being a multidimensional construct. However, instead of using the effectiveness–efficiency dichotomy to define these dimensions apriori, we use a number of performance indicators (encompassing financial, market, operations, and human resources areas) and use exploratory and confirmatory factor analysis to identify the factor structure of hospital performance.

Past studies have generally provided evidence that market orientation is related to organizational performance. One study related performance, measured in terms of return on assets, to top managers’ subjective assessments of market orientation for both commodity and non-commodity businesses [1]. Results showed that market orientation was an important determinant of performance although the nature of the relationship varied for the two types of businesses. Another study in the hospital industry showed that all three dimensions of market orientation (as identified by Kohli and Jaworski) were related to the financial performance of hospitals, although market intelligence and interfunctional coordination were more related to financial performance than organizational responsiveness [11]. In a prior published work the authors of the present article have also examined the relationship between market orientation and performance in the hospital industry [12]. Using a stepwise regression methodology it was found that market orientation was significantly related to performance. In addition, each performance dimension was affected by different dimensions of market orientation. These findings were interesting. However, only exploratory factor analysis was used in the measurement of market orientation and performance and the relationship of market orientation was examined with each performance dimension separately. In this paper we use the SEM approach to examine the dimensionality of market orientation and performance as well as the relationship between these constructs. The measurement models for market orientation and performance use confirmatory factor analysis to assess the goodness of fit of the dimensional structure of both market orientation and performance. As mentioned earlier, there is strong evidence in the literature that both market orientation and performance are multi-dimensional constructs. In the SEM, both market orientation and performance are therefore conceptualized as second-order factors with multiple indicators (dimensions) underlying these constructs. The SEM methodology is particularly useful in examining the causal relationships among multidimensional constructs. This study also examines the impact of hospital size on the market orientation–performance relationship. The existing literature on market orientation does not provide any evidence as to whether the market orientation–performance relationship is moderated by organization size. In the healthcare literature, hospital size, usually measured in terms of the number of beds, is recognized as an important determinant of various aspects such as the types of services offered and the cost per patient day [13]. Again, however, any impact of hospital size on the market orientation–performance relationship has not been examined in the healthcare literature. In this study we therefore examine the differences between large and small hospitals in terms of their market orientation, performance and the relationship between these constructs. While there is no a priori reason to believe that large and small hospitals would differ in either market orientation or performance, it is quite possible that the relationship between the constructs might be moderated

P.S. Raju et al. / Market orientation and performance in the hospital industry

by hospital size. In particular, one reasonable hypothesis would be that this relationship would be stronger for smaller hospitals. Since large hospitals are likely to have other significant advantages in terms of consumer awareness, accessibility to resources, and bargaining strength in the marketplace, it is possible that market orientation by itself might not affect performance for these hospitals as much as it would for smaller hospitals. The competitive strength of smaller hospitals, on the other hand, might be determined to a large degree by the extent of their market orientation and reflected in their performance. This hypothesis, however, must be considered exploratory at the present time since the literature does not provide us with a strong conceptual basis for making a connection between hospital size and the impact of market orientation on performance.

3. Methodology Data for this study were collected using a questionnaire that was mailed to the top executives of 740 hospitals in a five-state region (Kentucky, Minnesota, Mississippi, Ohio and Tennessee) in mid-western United States. This represented almost all the hospitals in the region (97%) and the five states accounted for approximately 12% of the hospitals in the nation. Usable responses were obtained from 175 hospitals for a response rate of 24%. In order to get an idea of how the sample compared with the population of hospitals the size distribution of hospitals in the sample was compared with the size distribution of all the hospitals in the region as well as the entire United States. Hospital size distribution in the sample based on the number of beds was: less than 100 beds, 26.7%; 100–200 beds, 43.75%; 300 or more beds, 29.5%. In both the five-state region as well as the US the comparable percentages were approximately 45%, 38% and 17%, respectively. It can be seen that the sample distribution was skewed toward the larger hospitals to some degree, probably because smaller hospitals are less likely or less willing to respond to surveys. Since the hospitals were not pre-selected on any particular ownership criterion or specialty, all types of hospitals were represented in the sample. Four surveys were mailed to the chief executive of each hospital. Instructions on the cover letter requested the chief executive to complete one survey and forward the other three surveys to three other senior executives of the hospital, preferably vice-presidents in the areas of quality, marketing, and operations. A total of 293 responses were received from the top executives of the 175 hospitals that responded. Approximately 37% of the hospitals sent in multiple responses while the rest sent in only a single response, usually from the CEO. Although the diversity of opinions among the top executives of hospitals is perhaps an interesting research topic in itself, that was not the focus of this study. Also, preliminary analysis revealed no major differences on relevant variables between hospitals that sent in a single response and those that sent in mul-

239

tiple responses. Therefore multiple responses from a particular hospital were averaged across the respondents from that hospital for each variable in order to come up with an aggregated response for that hospital. Such aggregation enabled the analysis to be conducted at the hospital level (with each hospital having the same weight) and derive implications for hospital strategy. Since a majority of the hospitals (63%) sent in only a single response any drawbacks of such aggregation were, hopefully, not serious. The survey instrument had questions relating to market orientation and performance, as well as other questions relating to general concerns in the hospital industry. Market orientation was measured with the MARKOR instrument [7]. However, since the original instrument had been developed within a manufacturing setting, the wording of the 30 original items was modified for use in a healthcare context. Performance can be measured using objective measures (such as ROI, market share, etc.) or judgmental measures which are based on executives’ perception of how the organization is performing relative to the competition. For cross-sectional studies judgmental measures would be better since these measures are more likely to reflect how an organization is performing at a particular point in time. Objective measures, on the other hand, often exhibit lagged effects since characteristics such as market orientation may not be reflected in bottom line measures for a considerable period of time. In addition, many organizations are often unwilling to reveal objective measures such as sales, profits, market share, etc. in surveys. Executives may have poor memory of such objective measures even if they wished to reveal them. On the other hand, most hospital executives would be able to give a subjective assessment of how the hospital is performing on various dimensions. Subjective assessments are especially likely to be more accurate when judgments have to be made with respect to the hospital’s competitors, whereas absolute objective measures might be more misleading. Judgmental measures have been used in the past by Jaworski and Kohli [14], Narver and Slater [1], and Kumar et al. [15] to measure business performance. Han et al. [2] used objective performance measures for banks but report that both objective and self-reported measures were not very different. For all the above reasons we opted to use judgmental measures in the present study. The hospital executives were asked to rate their hospitals on 19 performance variables relative to the competition. A scale of 1 (much worse than competition) to 5 (much better than competition) was used. The 19 performance variables were generated based on a review of hospital performance related literature as well as interviewing key executives at local hospitals. The analysis of the data essentially comprised of the following steps: 1. Exploratory factor analysis of the market orientation and performance variables separately in order to extract the dimensions of each construct.

240

P.S. Raju et al. / Market orientation and performance in the hospital industry

2. Confirmatory factor analysis (using SEM) of the market orientation and performance variables in order to determine if the extracted dimensions in step 1 offered a good fit to the data. Measurement models for market orientation and performance were tested at the individual construct level, aggregate level, and the second order level. 3. SEM in order to link market orientation with performance for all hospitals. 4. Separate SEM modeling for the small and large hospitals in order to examine if the path coefficient between market orientation and performance was different for the large and small hospitals. These steps are discussed in more detail in the next section. 4. Analysis and results 4.1. Exploratory factor analysis Exploratory factor analysis was conducted on the 30 market orientation items and yielded four factors. Details of this step are not discussed here since they are available in prior published work by the authors [12]. Only sixteen of the thirty items loaded on these four factors and, based on the items loading on each factor, the factors were labeled “intelligence generation” (factor 1), “customer satisfaction” (factor 2), “responsiveness to customers” (factor 3), and “responsiveness to competition” (factor 4). These sixteen items are shown as items M1 through M16 in table 1. In comparing this factor structure with other existing frameworks it can be seen that while the intelligence generation and responsiveness aspects were very similar to the dimensions postulated by Kohli and Jaworski [4] the emergence of customer and competitor oriented dimensions more closely resembles the Narver and Slater [1] framework. Thus the emerging factor structure appears to be a combination of the two existing frameworks. The customer satisfaction dimension, while it is somewhat unique and distinct from other frameworks, also appears to be quite relevant in a health service context. Since all four dimensions appear to have considerable face validity in a healthcare setting, it is quite possible that any differences from the earlier frameworks might reflect the differences between the manufacturing context and the healthcare service context. The convergent validity of the four dimensions was tested by correlating the factor scores on these dimensions with two separate measures of market orientation. One was a summated score on a subset of fifteen of the thirty original items that had been identified by Kohli et al. [7] as being a general market orientation factor. Although this is not an independent measure, each of the dimensions of market orientation would be expected to have a positive correlation with a general factor. The second measure was an independent single item subjective rating of the hospital’s marketing orientation as assessed by the respondents. Of

the eight correlations of the four dimensions with these two measures, seven were significant at the 0.01 level, demonstrating a fairly high convergent validity for the four factors extracted in this study. Cronbach’s alpha measures of reliability for the four factors were 0.82 for factor 1, 0.73 for factor 2, 0.69 for factor 3, and 0.71 for factor 4. All four values are either above or very close to the traditionally acceptable value of 0.70 in research [16]. Similar exploratory factor analysis was also conducted for the 19 performance measures. Three factors were extracted with thirteen of the nineteen measures loading on these factors. These thirteen performance variables are shown as items P1 through P13 in table 2. The factors were labeled “financial performance” (factor 1), “market/product development” (factor 2), and “internal quality” (factor 3). Financial performance comprises of variables such as net profit, profit to revenue ratio, return on investment, and cash flow from operations. Market/product development includes variables such as new product/service development, market development, the capacity to develop a unique competitive profile, and R&D aimed at new innovations. The third performance dimension “internal quality” is reflected in measures such as the quality of the service provided, mortality/morbidity rate, employee turnover, and cost per adjusted discharge. Cronbach’s alpha values for the three dimensions of performance were 0.95 for financial performance, 0.86 for product/market development, and 0.57 for internal quality. The reliability value for the internal quality dimension is somewhat lower than the usual acceptable value of 0.7. This is probably because this dimension comprises four items which are somewhat dissimilar in nature but are all related in some way to the quality of the work performed by the hospital. Some would prefer to include cost per adjusted discharge as a measure of financial performance. However, this variable loaded relatively highly (0.64) on the dimension of internal quality while its loading on the financial performance dimension was less than the specified cutoff of 0.4, and we chose to keep it as an internal quality measure pending the results of the confirmatory factor analysis. One option might have been to eliminate the internal quality dimension entirely due to its lower reliablility value. However, after careful consideration, we opted to retain this factor since it had an eigenvalue above 1.0, and the four items which loaded on this factor all had factor loadings above 0.6. 4.2. Structural equations modeling We now discuss the SEM methodology that was used to examine the market orientation and performance constructs and the relationship between them within the hospital industry. This methodology is attractive because it has the ability to deal with multiple relationships simultaneously while providing statistical efficiency [17]. This methodology also allows us to go from exploratory analysis to confirmatory analysis, where we can test or confirm

P.S. Raju et al. / Market orientation and performance in the hospital industry

241

Table 1 Measurement models for market orientation. Symbol

Description

Individual

Intelligence generation M1 In our hospital we meet with customers (i.e., physicians, businesses, insurance companies, and patients) at least once a year to find out what products or services they will need in the future. M2 Individuals from our operations interact directly with customers to learn how to serve them better. M3 In our hospital we do a lot of in-house research. M4 We survey customers at least once a year to assess the quality of our products and services. M5 We often talk with or survey those who can influence our patients, choices (e.g., physicians, health maintenance organizations). M6 We collect industry information through informal means (e.g., lunch with industry friends). Customer satisfaction M7 Data on customer satisfaction are disseminated at all levels in this hospital on a regular basis. M8 Customer complaints fall on deaf ears in this hospital. M9 When we find out that customers are unhappy with the quality of our service, we take corrective action immediately. M10 When we find that customers would like us to modify a product or service, the departments involved make concerted efforts to do so. Responsiveness to customers M11 We are slow to detect changes in our customers’ product/service preferences. M12 We are slow to detect fundamental shifts in our industry (e.g., competition, technology, regulation). M13 There is minimal communication between marketing and operations concerning market developments. M14 Our business plans are driven more by technological advances than by market research. Responsiveness to competition M15 If a major competitor were to launch an intensive campaign targeted at our customers, we would implement a response immediately. M16 We are quick to respond to significant changes in our competitors’ pricing structures.

Aggregate

Regression weight

T value

Regression weight

T value

0.65

6.31

0.67

6.41

0.60

5.97

0.66

6.38

0.60 0.67

5.98 6.44

0.60 0.63

6.04 6.23

0.75

6.88

0.74

6.86

0.60

–

0.58

–

0.52

6.35

0.53

6.52

0.51 0.79

6.14 –

0.51 0.77

6.33 –

0.84

8.49

0.85

9.73

0.68 0.66

5.76 –

0.65 0.70

6.68 –

0.63

5.67

0.62

6.45

0.45

4.53

0.45

4.94

0.87

–

0.61

–

0.53

5.11

0.75

5.19

– Fixed for estimation.

a prespecified relationship. The three software packages commonly used for SEM are AMOS, LISREL and EQS. In this paper we have used AMOS which is part of the SPSS software package [18]. Data from all 175 hospitals were used for the SEM analysis. Occasional missing data on variables was handled by replacing them with the mean value. The percentage of missing data in the market orientation and performance variables across the 175 hospitals was calculated to be 1.8%. Correlations among the relevant variables constituted the input for the various SEM analyses in this paper. Tables 1 and 2 summarize the measurement model results for market orientation and performance, respectively. Each table shows the standardized regression weight for each variable at both the individual construct level as well as the aggregate level. The individual construct level weights relate to the separate analysis performed for each factor including just the variables that related to that specific factor. The aggregate model includes all the factors (latent variables) and examines the relationships among the factors for the mar-

ket orientation and performance constructs. For example, in the case of the market orientation construct, customer satisfaction, intelligence generation, responsiveness to customers, and responsiveness to competition were all included in the model along with the variables that load (indicators) on each factor. The standardized regression weights for all the variables are highly significant (as indicated by the t-values) for both the category level analysis and the aggregate analysis. Several goodness of fit measures corresponding to these models are summarized in table 4. Examination of these goodness-of-fit measures shows that they are uniformly good, indicating that the measurement models fit the data well. The minimum discrepancy (χ2 /df) is less than 2.0 in all cases (it should be between 0 and 5 for a good fit with lower values indicating a better fit), the root mean residual (RMR) values are very low (a value of 0 indicates perfect fit), and the goodness of fit (GFI) and adjusted goodness of fit (AGFI) indexes are both close to 1.0 (a value of 1.0 indicates perfect fit). Only the Hoelter index (for which the critical N > 200 for a good fit) gives somewhat mixed

242

P.S. Raju et al. / Market orientation and performance in the hospital industry Table 2 Measurement models for performance. Symbol

Description

Individual

Aggregate

Regression weight

T value

Regression weight

T value

Financial performance P1 Net profits. P2 Return on investment. P3 Cash flow from operations. P4 Return on assets. P5 Profit to revenue ratio.

0.94 0.92 0.82 0.82 0.88

– 22.21 16.29 16.03 19.57

0.90 0.88 0.87 0.83 0.92

– 22.51 16.84 15.22 18.99

Market/product development P6 New product/service development. P7 Investments in R&D aimed at new innovation. P8 Capacity to develop a unique competitive profile. P9 Market development.

0.87 0.76 0.75 0.76

– 10.94 10.88 11.04

0.79 0.68 0.79 0.81

– 10.77 10.65 10.97

Internal quality P10 Mortality and morbidity rate. P11 Service quality as perceived by customers. P12 Cost per adjusted discharge. P13 Employee turnover.

0.45 0.79 0.29 0.55

– 3.80 2.85 4.23

0.40 0.78 0.37 0.54

– 4.70 3.44 4.22

– Fixed for estimation.

results. However, the Hoelter index is not universally accepted as it can lead to an overly pessimistic assessment of fit, especially for smaller samples [19]. It should also be noted that in the case of the construct level analysis for the fourth factor for market orientation, i.e., responsiveness to competition, the fit is perfect because there were only two variables that load on the factor and consequently there are only two variables in the analysis. In general, the measurement models strongly support the dimensionality and structure of the market orientation and performance constructs as identified by the exploratory factor analysis. In addition to the aggregate model results shown in tables 1, 2 and 4 we also tested second order measurement models for both market orientation and performance. This analysis tested if, in fact, the four factors of market orientation could be combined into the single construct of market orientation and the three performance factors could similarly be combined into the single construct of performance. In order to conserve space the detailed results are not reproduced here. Once again, the standardized regression weights were all statistically significant as indicated by the t-values. The goodness-of-fit measures also consistently indicated an extremely good fit of the market orientation and performance models to the data. The analysis showed that the squared multiple correlations between market orientation and the four factors (i.e., reliabilities of the four factors or indicators) were 0.54 for intelligence generation, 0.49 for customer satisfaction, 0.67 for responsiveness to customers, and 0.56 for responsiveness to competitors. These figures indicate the percentage of variance in each factor that is explained by the market orientation construct (for example, 54% of the variance in intelligence generation among hospitals can be attributed to differences in their market orientation. In the case of performance, the squared correlation

coefficients for the three factors were 0.42 for financial performance, 0.99 for product/market development, and 0.67 for internal quality. Once again, the high reliablility of the internal quality factor supports the decision to keep the cost per adjusted discharge as part of this factor. Perhaps, from the viewpoint of a hospital executive, the cost per adjusted discharge is a cost measure that is an indicator of quality to a greater degree than it is an indicator of financial performance (typically revenue measures). Overall, the analysis showed that a great degree of the variance in the four factors of market orientation and the three factors of performance is encompassed within the constructs of market orientation and performance. This shows that the latent variables of market orientation and performance can be reliably operationalized using the dimensions that were specified for each construct and therefore validates the factor structure of the market orientation and performance constructs. However, one should be somewhat cautious in placing complete confidence in the factor structures. Although commonly done, some authors do not recommend using the same data for specifying a model and for assessing the fit of the data to the model [20,21]. Since both market orientation and performance were measured with self-reported instruments it was also necessary to establish that these were in fact separate constructs and that the instruments were not essentially measuring the same construct. Otherwise, one could hypothesize that any relationship found between these constructs was a function of how they were measured. From a conceptual viewpoint, such a hypothesis is untenable since the market orientation concept is well established and tested in the marketing literature (as outlined earlier). It is therefore highly unlikely that this construct would merely be a surrogate measure of performance. However, in order to empirically establish the

P.S. Raju et al. / Market orientation and performance in the hospital industry Table 3 Path model: relationship between market orientation and performance. Constructs Performance – Market orientation (MO) Intelligence generation – (MO) Customer satisfaction – (MO) Customer responsiveness – (MO) Competitor responsiveness – (MO) Financial performance – performance Market development – performance Internal quality – performance

Regression weights

T values

0.80 0.63 0.61 0.68 0.66 0.68 0.84 0.72

6.00 – 6.32 6.80 6.67 – 8.70 7.99

– Fixed for estimation.

discriminant validity of the two constructs a measurement model was tested in which the four market orientation dimensions and the three performance dimensions were used as indicators of one unifying construct. This model indicated a very poor fit to the data. The adjusted goodness-offit index for this model was only 0.55 and the χ2 /df ratio was 12.5, far above the maximum acceptable value of 5. Discriminant validity was therefore confirmed for the two constructs. The final step in the analysis was the path model that linked market orientation with performance. Table 3 summarizes the standardized regression weights and t-values for this analysis. The standardized regression weight between market orientation and performance is quite high (0.80) and the t-value is statistically significant. All other standardized regression weights are also highly significant. The goodness-of-fit measures for this model in table 4 again show a very good fit of the model to the data. Only the RMR value of 0.32 is a little high, but this by itself is inconclusive since the regression weight between market orientation and performance is high and the other goodness-of-fit indices are quite acceptable. There are various reasons why residuals are sometimes large [19, p. 257] and the residuals also tend to get larger when the sample size gets smaller. Figure 1 shows the path analysis model. It should be noted in figure 1 that the individual market orientation and performance variables were not used in the analysis. Instead, the four market orientation factors and the three performance factors were represented by the summated scores of the items loading on each factor. This was done in order to conserve the degrees of freedom since the sample size was not large enough to accommodate all the individual variables into the analysis. Figure 1 shows the path coefficient between market orientation and performance to be 0.80, which is quite strong. The coefficient of determination, which is analogous to a squared multiple correlation coefficient, and shows the proportion of variance in the endogenous variable (performance) that can be accounted for by the exogenous variable (market orientation) can therefore be calculated as 0.64 [17, p. 703]. A considerable degree of the variance in hospital performance can therefore be explained by market orientation. In order to examine the impact of hospital size on market orientation, performance, and the relationship between

243

the two constructs we used a median split to break up the hospitals into two groups, small hospitals (N = 88, 6184 beds) and large hospitals (N = 87, >184 beds). We first examined the mean values for the four market orientation dimensions and three performance dimensions for the two groups. Summated scores of the indicator variables were used to represent each dimension. The detailed results for this analysis are not presented here since none of the dimensions of market orientation or performance was significantly different for the two groups. This indicates that large and small hospitals do not inherently differ in the degree to which they are market oriented or how well they perform with respect to their competitors. We then performed separate path analyses for the two sets of hospitals. The path analysis for the small hospitals is shown in figure 2 and the analysis for the large hospitals is shown in figure 3. Table 4 summarizes the goodness-of-fit measures for these models also. It can be noted that the goodness-of-fit measures are quite acceptable for both models. This shows that market orientation is strongly related to performance for both small and large hospitals. Only the RMR values and the Hoelter index are somewhat questionable, perhaps due to the smaller sample sizes for the groups. When the sample was split in two the RMR value increased to 0.40 (small hospitals) and 0.42 (large hospitals) from the value of 0.32 for the combined sample. This is in conformity with the stated limitation of RMR as being dependent on sample size. Bollen [19, p. 257] states that large residuals can reflect poor parameter estimation. However, there can be two other reasons why residuals can be large. The first is the scales of the observed variables. When the observed variables are measured in different units a larger residual can result for a variable when its range is larger than that of other variables. That is not the case in this paper since we used the correlation matrix among variables for the analysis. The second reason is small sample size, which is more likely to be the reason here. Bollen [19] also states that the correlation residual has a theoretical range from −2 to +2. Even though the RMR values in this analysis are not very close to 0, they also are not however anywhere near the extreme limits for indicating a poor fit. The most interesting aspect of the results, however, is that the standardized regression coefficient between market orientation and performance in figure 2 (0.92) for the small hospitals appears to be much higher than the regression coefficient in figure 3 (0.68) for the large hospitals. The percentage of variance in performance explained by market orientation (R2 ) is therefore much higher for smaller hospitals (85%) as compared to the larger hospitals (46%). In order to further confirm whether the relationship between market orientation and performance is really different between large and small hospitals we conducted a multigroup analysis [22]. This analysis examines if the loadings (the lambda or LX matrix), error variances (the theta-delta or TD matrix), and covariance between the latent constructs (the PHI matrix) is the same for the two

244

P.S. Raju et al. / Market orientation and performance in the hospital industry Table 4 Goodness-of-fit statisticsa χ2 /df

RMR

GFI

AGFI

Hoelter∗

6.64 1.66 0.70 –

0.83 0.83 0.35 –

0.02 0.02 0.01 –

0.99 0.99 0.99 1.00

0.97 0.98 0.99 –

407 628 1480 –

5.98 1.99 3.72

1.50 0.99 1.86

0.01 0.01 0.02

0.99 0.99 0.99

0.95 0.97 0.94

276 524 280

Aggregate analysis Aggregate model (market orientation) Aggregate model (performance)

136.26 98.03

1.40 1.63

0.05 0.04

0.91 0.92

0.88 0.88

155 141

Relationship between market orientation and performance model

17.84

1.37

0.32

0.97

0.94

219

11.64

0.90

0.40

0.97

0.92

168

14.97

1.15

0.42

0.95

0.90

129

Variable

χ2

Market orientation Intelligence generation Customer satisfaction Responsiveness to customers Responsiveness to competitors Performance Financial performance Market development Internal quality

Market orientation and performance model (small hospitals) sample size = 88 Market orientation and performance (large hospitals) sample size = 87 a Sample

size = 175. 6 0.05. – Cannot be calculated. Chi-square is the likelihood ratio, or the likelihood that a specific model represents the causal relations among observed variables. A smaller value of chi-square represents a better fit. The chi-square-to-degree-of-freedom index is a standardized measure, with a smaller value also representing a better fit. GFI is the goodness-of-fit index. The RMR (root mean square residual) is the square root of the average squared amount by which the sample variances and covariances differ from their estimates obtained under the assumption your model is correct. The smaller the RMR is, the better. An RMR of zero indicates a perfect fit. AGFI, adjusted goodness-of-fit index, takes into account the degrees of freedom available for testing the model. The AGFI is bounded above by one, which indicates a perfect fit. It is not, however, bounded below by zero. Hoelter’s (1983) “critical N ” is the largest sample size for which one would accept the hypothesis that a model is correct.

*α

groups. If the chi-square value changes by a significant amount when these parameters are constrained to be equal between the groups, as compared to when they are unconstrained, then the factor structure in the two groups is different which might account for the difference found in the relationship. A non-significant change in the chi-square value, on the other hand, would imply that the factor structure was the same for the two groups and that the difference in relationship found cannot therefore be attributed to any difference in factor structure between the groups. For the unconstrained model, the χ2 value was 26.61 with 26 degrees of freedom. This is the same as the sum of the χ2 values reported for the two separate models for the large and small hospitals in table 4. In the constrained model the χ2 value was found to be 37.13 with 34 degrees of freedom. The difference in χ2 between the two models is 10.50 with 8 degrees of freedom which is not statistically significant at an alpha of 0.05. Thus, it can be concluded that the factor structures for large and small hospitals are the same and that any difference in factor structure cannot therefore account for the difference in the market orientation–performance relationship found for the two groups. The fact that the relationship between market orientation and performance is different for large and small hospitals,

especially when there seems to be no inherent difference in the degree of market orientation or performance levels of the two types of hospitals, has some major ramifications for hospitals. These implications will be addressed later in this paper.

5. Discussion The results of the SEM modeling illustrate many interesting details of the relationship between market orientation and performance in a hospital industry setting. Our exploratory factor analysis of the 30 modified items of the MARKOR instrument developed by Kohli et al. [7] indicated that, within the health care setting, market orientation could be characterized by four dimensions – intelligence generation, customer satisfaction, responsiveness to customers, and responsiveness to competition. These factors are intuitively appealing, but the SEM methodology enabled us to confirm and validate the four factor structure for market orientation with three different levels of analyses (the construct level, the aggregate level, and the second order analysis).

P.S. Raju et al. / Market orientation and performance in the hospital industry

245

Figure 1. Market orientation and performance.

Figure 2. Market orientation and performance for small hospitals.

Similarly, in the case of performance, the exploratory factor analysis indicated the presence of three factors – financial performance, market/product development, and internal quality. Again, the SEM analysis strongly supported the three-factor structure. Our analysis also confirmed discriminant validity for the market orientation and performance constructs and showed that they are not just different measures of a common construct. Finally, the path analysis model produced a very good fit with the data and showed that almost 64% of the variance in hospital performance could be explained by the market orientation of hospitals. Further analysis showed that this relationship was much stronger for the smaller hospitals with 85% of the variation in performance explained by market orientation, as compared to the larger hospitals (only about 46% of the variance explained). Interestingly, the small and large hospitals did not differ in the degree of market orientation or performance as indicated by the mean values of individual dimensions for both constructs.

industry. The results clearly reinforce the notion that market orientation is a multi-dimensional construct. Many researchers and practitioners have a tendency to define and/or measure market orientation as an uni-dimensional construct. Quite often, the dimension of customer satisfaction is the one that is predominant within hospitals as is evidenced by the many surveys that hospitals often undertake to measure patient satisfaction and other similar customer related outcomes. According to this study, such an approach might not be fruitful. In order to understand whether a hospital is truly market oriented one has to focus on all four dimensions of market orientation – intelligence generation, customer satisfaction, responsiveness to customers, and responsiveness to competition. Intelligence generation, responsiveness to customers, and responsiveness to competition are aspects which would most likely be determined by policies set by upper management. These three dimensions should not be ignored, or given less prominence than customer satisfaction, when assessing whether a hospital is market oriented. Similar arguments can be made for the performance dimension. Most organizations have a tendency to interpret performance purely in terms of bottom-line financial measures. This is a short-sighted approach, and this study shows that market/product development and internal quality are two additional dimensions of performance that should

6. Implications The results of this study have some important implications for both researchers and practitioners in the hospital

246

P.S. Raju et al. / Market orientation and performance in the hospital industry

Figure 3. Market orientation and performance for large hospitals.

not be ignored. Hence, performance should also be interpreted as a multi-dimensional construct. Assessment of hospital performance could therefore be much more complex than it appears. The strong relationship between market orientation and performance has important ramifications for hospital executives. The fact that, on the average, almost two-thirds of the variance in performance can be accounted for by market orientation implies that “market orientation” should be part of the standard vocabulary within hospitals. This construct is not well understood and most hospitals have a tendency to emphasize only certain aspects of marketing such as public relations, advertising, or patient satisfaction. Market orientation, on the other hand, has something to do with the entire culture within the organization and emphasizes how information is obtained, disseminated, and used to better understand and be responsive to customers and competitors. The differences in the results for small and large hospitals found in this study are of particular interest to hospital administrators. First, the finding that the large and small hospitals are not significantly different in terms of their market orientation or performance levels implies that hospital size does not help or hinder these aspects. Thus, smaller hospitals can be just as market oriented and perform just

as well as large hospitals and vice versa. Ultimately, it is therefore up to each hospital whether it desires to be more market oriented or perform better than the competition. The most interesting result in this paper is that the relationship between market orientation and performance is very different for large and small hospitals. For smaller hospitals, our study shows that market orientation has a tremendous influence on performance, with almost 85% of the variance in performance being attributed to market orientation. Thus, being market oriented could be vital for the survival of smaller hospitals. This is especially significant since smaller hospitals are the ones that are perhaps more likely to put less emphasis on marketing in general. These hospitals are likely to have greater financial pressures, and any marketing that is done is more likely to be ad hoc and not follow any conceptual framework. For the larger hospitals, on the other hand, less than 50% of the variance in performance can be attributed to market orientation. Although this is still a substantial effect, it means that it might not be as critical for larger hospitals to be market oriented. Larger hospitals usually tend to have certain advantages in the marketplace such as access to resources, bargaining capability, economies of scale, and a higher awareness among customers. At the same time, the substantial impact of market orientation on performance, even for the larger hospitals, shows that these hospitals cannot really afford to ignore being market oriented. Interestingly, even though market orientation is not as critical for the larger hospitals, these hospitals might have to work harder in order to become market oriented. This is because most of the dimensions of market orientation can be accomplished more easily within smaller organizations. Smaller hospitals might find it easier to disseminate market information within the organization and be more responsive to customers and competitors. Thus, market orientation is an important concept for both large and small hospitals. While large hospitals might benefit more by understanding how market orientation interacts with other factors to determine performance, smaller hospitals might gain more by putting a high degree of emphasis on market orientation since that appears to be the predominant determinant of performance. For researchers, the findings have several implications. As mentioned earlier in this paper, several researchers have suggested that market orientation and performance are related. However, we are not aware of any studies that specifically suggest that this relationship is moderated by organization size. Hence, this finding is particularly significant and needs to be examined further in future studies. In fact, it appears that the research topic of organization size and its relationship to market orientation and performance could be a very fruitful area for future research. The measurement of both market orientation and performance are also areas that require research attention. This study suggests that, even though market orientation and performance are multi-dimensional measures, they could be measured using fairly simple instruments. In this study the four factors of market orientation incorporated only 16

P.S. Raju et al. / Market orientation and performance in the hospital industry

of the original 30 items in the MARKOR instrument and the three factors of performance incorporated only 13 of the 19 judgmental measures. The simplicity of such measures could be a major advantage to researchers, especially those using surveys for data collection. However, whether the same items could be used across different industries or if the instruments would need to be adapted for various industries is an issue that requires further study. The comparison of different instruments for measuring market orientation and performance would also be useful. As we have pointed out in this paper, researchers have defined market orientation in several ways. It is not clear, therefore, if instruments based on these different definitions would yield different results. Other ways of operationalizing and measuring performance also need to be explored. This paper used a judgmental measure and revealed three dimensions of performance, namely financial performance, market/product development, and internal quality. In the healthcare TQM literature, researchers have operationalized performance in terms of effectiveness and efficiency measures. Perhaps the use of different operationalizations of performance might yield other indicators or dimensions of performance and enable us to fully explore the impact of market orientation on organizational performance. One final consideration is that, in this study, both market orientation and performance were measured using judgmental (perceptual) measures. It would be useful to compare results using both objective and subjective measures since each has relative advantages and disadvantages. As mentioned earlier, objective measures are sometimes difficult to obtain since organizations are reluctant to provide them (especially financial measures) but it might be useful to obtain them whenever possible. In conclusion, this study has found strong evidence for the multi-dimensionality of both market orientation and performance within the hospital industry. Also, there is a strong relationship between market orientation and hospital performance. This relationship is moderated, however, by the size of the hospital with smaller hospitals exhibiting a much stronger relationship. Understanding this relationship and utilizing it for strategic purposes might very well turn out to be the key to differentiating the winners from the losers in the intensely competitive hospital industry of the future. Acknowledgement The authors are very grateful to Mr. Gerald T. Pierce, former Executive Director, Center for Excellence for Quality, Kentucky State University, Frankfort, KY 40601 for his help in the data collection phase of this study.

247

References [1] J.C. Narver and S.F. Slater, The effect of market orientation on business profitability, Journal of Marketing 54 (Oct. 1990) 20–35. [2] J.K. Han, N. Kim and R. Srivastava, Marketing orientation and organizational performance: is innovation a missing link?, Journal of Marketing 62 (Oct. 1998) 30–45. [3] R.F. Lusch and G.R. Laczniak, The evolving marketing concept, competitive intensity, and organizational performance, Journal of the Academy of Marketing Science 15 (Fall 1987) 1–11. [4] A.K. Kohli and B.J. Jaworski, Market orientation: the construct, research propositions, and managerial implications, Journal of Marketing 54 (April 1990) 1–18. [5] B.P. Shapiro, What the hell is ‘market oriented’?, Harvard Business Review 66 (Nov.–Dec. 1988) 119–125. [6] P. Kotler and R.N. Clarke, Marketing for Health Care Organizations (Prentice-Hall, Englewood Cliffs, NJ, 1987). [7] A.K. Kohli, B.J. Jaworski and A. Kumar, MARKOR: a measure of market orientation, Journal of Marketing Research 30 (Nov. 1993) 467–477. [8] S.M. Shortell, J.L. O’Brien, J.M. Carman and R.W. Foster, Assessing the impact of continuous quality improvement/total quality management: concept versus implementation, Health Services Research 30(2) (1995) 377–401. [9] M.A. Counte, G.L. Glandon, D.M. Oleske and J.P. Hill, Improving hospital performance: issues in assessing the impact of TQM activities, Hospital and Health Services Administration 40(1) (1995) 80–94. [10] W.R. Scott and S.M. Shortell, Organizational performance: managing for efficiency and effectiveness, in: Health Care Management: A Text in Organization Theory and Behavior, eds. S.M. Shortell and A.D. Kaluzny (Wiley, New York, 1983). [11] D.R. McDermott, F.J. Franzak and M.W. Little, Does marketing relate to hospital profitability?, Journal of Health Care Marketing 13(2) (1993) 18–25. [12] P.S. Raju, S.C. Lonial and Y.P. Gupta, Market orientation and performance in the hospital industry, Journal of Healthcare Marketing 15(4) (1995) 34–41. [13] F.D. Wolinsky, The Sociology of Health: Principles, Practitioners, and Issues, 2nd ed. (Wadsworth, Belmont, CA, 1988). [14] B.J. Jaworksi and A.K. Kohli, Market orientation: antecedents and consequences, Journal of Marketing 57 (July 1993) 53–70. [15] K. Kumar, R. Subramanian and C. Yauger, Performance-oriented: toward a successful strategy, Marketing Health Services (Summer 1997) 10–20. [16] J.C. Nunnally, Psychometric Theory, 2nd ed. (McGraw-Hill, New York, 1978). [17] J.F. Hair, R.E. Anderson, R.L. Tatham and W.C. Black, Multivariate Data Analysis, 4th ed. (Prentice-Hall, NJ, 1995). [18] J.L. Arbuckle, Amos User’s Guide, Version 3.6 (Small Waters Corp., Chicago, IL, 1997). [19] K.A. Bollen, Structural Equations with Latent Variables (Wiley, New York, 1989). [20] Leamer, Specification Searches (Wiley, New York, 1978). [21] J.S. Long, Confirmatory Factor Analysis (Sage, Beverly Hills, 1983). [22] S. Sharma, Applied Multivariate Techniques (Wiley, New York, 1996).